Lab Meeting 10/14, 2009 (Shao-Chen): Distributed Multirobot Localization(IEEE Transactions on Robotics and Automation Oct. 2002)

Title: Distributed Multirobot Localization(IEEE Transactions on Robotics and Automation Oct. 2002)

Authors: Stergios I. Roumeliotis,  George A. Bekey

Abstract:

In this paper, we present a new approach to the
problem of simultaneously localizing a group of mobile robots
capable of sensing one another. Each of the robots collects sensor
data regarding its own motion and shares this information with
the rest of the team during the update cycles. A single estimator,
in the form of a Kalman filter, processes the available positioning
information  from  all  the  members  of  the  team  and  produces
a pose estimate for every one of them. The equations for this
centralized  estimator  can  be  written  in  a  decentralized  form,
therefore  allowing this  single  Kalman  filter  to  be  decomposed
into a number of smaller communicating filters. Each of these
filters processes sensor data collected by its host robot. Exchange
of information between the individual filters is necessary only
when two robots detect each other and measure their relative
pose. The resulting decentralized estimation schema, which we
call collective localization, constitutes a unique means for fusing
measurements collected from a variety of sensors with minimal
communication  and  processing  requirements.  The  distributed
localization algorithm is applied to a group of three robots and
the improvement in localization accuracy is presented. Finally, a
comparison to the equivalent decentralized information filter is
provided.

Link

CMU talk: The Inner Workings of face Processing: From human to computer perception and back

CMU VASC Seminar
Monday, October 12
2pm-3pm

The Inner Workings of face Processing: From human to computer perception and back
Aleix Martinez
Ohio State University

Abstract:
Faces and emotions shape our daily life in many different ways. We are so intrigued about such effects that writers, poets and painters have been depicting and portraying them for centuries. Why does the male character in Wood's "American Gothic" seem sad? Why do kids elongate their faces when they are upset? Why do some people always seem angry? Why do we recognize identity from faces so easily? Why is it so hard to learn non-manuals (i.e., facial expressions of grammar) in sign languages, when native signers do this effortlessly? In short, what are the dimensions of our computational (cognitive) space responsible for these face processing tasks? If we are to understand why things appear as they do and how cognitive disorders in Autism, schizophrenia and Huntington’s disease develop, we need to define how the brain codes and interpreters faces, emotions and grammar. This is also important for the design of technology – as devices need to interact with us. In this talk, I will outline the research framework I use to study face perception and the related topics of emotion and grammar. This consists of a multidisciplinary approach in cognitive science, including psychophysical studies and the design of computer algorithms for the analysis of face images. We will review the big questions about this computational space. In doing so, we will see that the ability of human observers to process face images is truly remarkable, suggesting that some abstract, yet simple representation that is unaffected by a large number of image transformations is at work. We will summarize our current understanding of this representation.

Bio:
Aleix M. Martinez is an Associate Professor in the Department of Electrical and Computer Engineering at The Ohio State University (OSU), where he is the founder and director of the Computational Biology and Cognitive Science Lab. He is also affiliated with the Department of Biomedical Engineering and to the Center for Cognitive Science. Prior to joining OSU, he was affiliated with the Electrical and Computer Engineering Department at Purdue University and with the Sony Computer Science Lab. He serves as an associate editor of IEEE Transactions on Pattern Analysis and Machine Intelligence and of Image and Vision Computing and has been an area chair of CVPR. Aleix has spent his time wondering why he is such a bad face recognizer and why people attribute social labels to faces of unknown individuals. His other areas of interest are learning, vision and linguistics.

Lab Meeting 10/6, 2009 (Casey): Towards Visual Localization, Mapping and Moving Objects Tracking by a Mobile Robot: a Geometric and Probabilistic ...

Title: Towards Visual Localization, Mapping and Moving Objects Tracking by a Mobile Robot: a Geometric and Probabilistic Approach

Author: Joan Solà

(PhD Thesis, 2007/02)

(LAAS-CNRS laboratory in Toulouse, Occitania, France )

Outline:

1. Undelayed initialization in monocular SLAM

2. Binocular SLAM

3. Moving object detection and tracking

[PhD Thesis pdf file]

[Joan Solà's website]

NTU talk: Current Challenges in Vision-Based Driver Assistance Systems

Title: Current Challenges in Vision-Based Driver Assistance Systems
Speaker: Prof. Reinhard Klette, The University of Auckland, Tamaki campus
Time: 10:00am, Sep 28 (Mon), 2009
Place: Room 105, CSIE building

Abstract: The talk starts with informing briefly about the .enpeda..project at The University of Auckland, and goals in vision-based driver assistance systems (DAS) in general, illustrated by accident statistics. Lane and corridor detection is a traditional DAS subject, and curved and unmarked roads define still a challenge. A solution for corridor (i.e., the expected space to drive in) detection is discussed based on applying the Euclidean distance transform. The main part of the talk is then about current stereo and optic flow algorithms on real-world (stereo) sequences. Prediction error analysis and evaluations on synthetic DAS sequences are discussed as possible options, and conclusions are drawn, such as the suggestion that correspondence algorithms should use residual images as input rather than the original sequences. Finally, a performance evaluation approach is illustrated which is currently under implementation, using a 3D model of a real scene for generating real-world sequences with ground truth for stereo and optical flow.

Short Biography: See Dr. Klette's resaerch at
http://www.mi.auckland.ac.nz/index.php?option=com_content&view=article&id=57&Itemid=49

Lab Meeting September 23rd, 2009 (Jeff): Topological Modeling and Classification in Home environment using Sonar gridmap

Title: Topological Modeling and Classification in Home environment using Sonar gridmap

Authors: Jinwoo Choi, Minyong Choi, Kyoungmin Lee and Wan Kyun Chung

Abstract:

This paper presents a method of topological representation and classification in home environment using only low-cost sonar sensors. Approximate cell decomposition and normalized graph cut are applied to sonar gridmap to extract graphical model of the environment. The extracted model represents spatial relation of the environment appropriately by segmenting several subregions. Moreover, node classification is achieved by applying template matching method to a local gridmap. Rotational invariant matching is used to obtain candidate location for each node and the true node can be classified by considering detail distance information. The proposed method extracts well-structured topological model of the environment and classification also results in reliable matching even under the uncertain and sparse sonar data. Experimental results verify the performance of proposed environmental modeling and classification in real home environment.

Link:
ICRA2009
http://pal.csie.ntu.edu.tw/pub/Conferences/2009_ICRA/Conference/data/papers/1720.pdf

Computerized Face-Recognition Technology Is Still Easily Foiled by Cosmetic Surgery

Computerized Face-Recognition Technology Is Still Easily Foiled by Cosmetic Surgery

In the first test of face-recognition technology vs. cosmetic surgery, face recognition loses.

BY Willie D. Jones // September 2009

For years, developers of face-recognition algorithms have been battling the effects of awkward poses, facial expressions, and disguises like hats, wigs, and fake moustaches. They’ve had some success, but they may be meeting their match in plastic surgery.

Systematic studies have tested face-recognition algorithms in a variety of challenging situations—bad lighting, for example—”but none of those conditions had nearly the effect of plastic surgery,” says Afzel Noore, a computer science and electrical engineering professor at West Virginia University, in Morgantown. In June, Noore reported the results of the first experimental study to quantify the effect of plastic surgery on face-recognition systems, at the IEEE Computer Society’s Computer Vision and Pattern Recognition conference, in Miami. His team of collaborators is based in West Virginia and at the Indraprastha Institute of Information Technology, Delhi, in India.

Using a database containing before-and-after images from 506 plastic surgery patients, Noore and his colleagues tested six of the most widely used face-recognition algorithms. Even in pictures where the subject was facing forward and the lighting was ideal, the best of the algorithms matched a person’s pre- and postsurgery images no more than about 40 percent of the time. The researchers found that for local alterations—say, a nose job, getting rid of a double chin, or removing the wrinkles around the eyes—today’s systems could make a match roughly one-third of the time. For more global changes like a face-lift, the results were dismal: a match rate of just 2 percent.

”We have to devise systems for security applications knowing that people will aim to circumvent them,” says Noore. In particular, researchers must examine a further complication of the plastic surgery problem—the compounding effects of a series of surgeries over time.

Meanwhile, Noore and his coauthors are testing a game-changing hypothesis: that even after plastic surgery, there are features beneath the skin but still observable that remain unchanged.

CMU PhD Thesis: Spectral Matching

Spectral Matching, Learning, and Inference for Computer Vision

Marius Leordeanu

doctoral dissertation, tech. report CMU-RI-TR-09-27, Robotics Institute, Carnegie Mellon University, July, 2009

Abstract: Several important applications in computer vision, such as 2D and 3D object matching, object category and action recognition, object category discovery, and texture discovery and analysis, require the ability to match features efficiently in the presence of background clutter and occlusion. In order to improve matching robustness and accuracy it is important to take in consideration not only the local appearance of features but also the higher-order geometry and appearance of groups of features. In this thesis we propose several efficient algorithms for solving this task, based on a general quadratic programming formulation that generalizes the classical graph matching problem. First, we introduce spectral graph matching, which is an efficient method for matching features using both local, first-order information, as well as pairwise interactions between the features. We study the theoretical properties of spectral matching in detail and show efficient ways of using it for current computer vision applications. We also propose an efficient procedure with important theoretical properties for the final step of obtaining a discrete solution from the continuous one. We show that this discretization step, which has not been studied previously in the literature, is of crucial importance for good performance. We demonstrate its efficiency by showing that it dramatically improves the performance of state-of-the art algorithms. We also propose, for the first time, methods for learning graph matching in both supervised and unsupervised fashions. Furthermore, we study the connections between graph matching and the MAP inference problem in graphical models, for which we propose novel inference and learning algorithms. In the last part of the thesis we present an application of our matching algorithm to the problem of object category recognition, and a novel algorithm for grouping image pixels/features that can be effectively used for object category segmentation.

Link: WWW, PDF

Talk: VASC Seminar: Jason Saragih Face Alignment through Subspace Constrained Mean-Shifts

Title:Face Alignment through Subspace Constrained Mean-Shifts

Author:Jason Saragih Post-Doc, Robotics, CMU

September 14, 2009, 2:30pm-3:00pm, NSH 3305

Abstract

Deformable model fitting has been actively pursued in the computer vision community for over a decade. As a result, numerous approaches have been proposed with varying degrees of success. A class of approaches that has shown substantial promise is one that makes independent predictions regarding locations of the model’s landmarks, which are combined by enforcing a prior over their joint motion. A common theme in innovations to this approach is the replacement of the distribution of probable landmark locations, obtained from each local detector, with simpler parametric forms. This simplification substitutes the true objective with a smoothed version of itself, reducing sensitivity to local minima and outlying detections. In this work, a principled optimization strategy is proposed where a nonparametric representation of the landmark distributions is maximized within a hierarchy of smoothed estimates. The resulting update equations are reminiscent of mean-shift but with a subspace constraint placed on the shape’s variability. This approach is shown to outperform other existing methods on the task of generic face fitting.

Speaker Biography

Jason Saragih joined the Robotics Institute as a Post-doctoral Fellow in 2008. He received both his BEng and PhD from the Australian National University in 2004 and 2008 respectively. His research interests concern the modeling and registration of deformable models.

CMU PhD Thesis Proposal: Robust Monocular Vision-based Navigation for a Miniature Fixed-Wing Aircraft

PhD Thesis Proposal
Myung Hwangbo
Carnegie Mellon University

September 15, 2009, 1:00 p.m., Newell Simon Hall 1109

Title:
Robust Monocular Vision-based Navigation for a Miniature Fixed-Wing Aircraft

Abstract:
Recently the operation of unmanned aerial vehicles (UAVs) has expanded from military to civilian applications. Contrary to remote-controlled tasks in a high altitude, low-altitude flight in an urban environment requires a higher level of autonomy to respond to complex and unpredictable situations. Vision-based methods for autonomous navigation have been a promising approach because of multi-layered information delivered by images but their robustness in various situations has been hard to achieve. We propose a series of monocular computer vision algorithms combined with vehicle dynamics and other navigational sensors in GPS-denied environments like an urban canyon. We use a fixed-wing model airplane of 1m wing span as our UAV platform. Because of its small payload and limited communication bandwidth to off-body processors, particular attention is paid to both realtime and robustness at every level of vision processing of low-grade images.

In point-to-point navigation, state estimation is based on the structure-from-motion method (SFM) using natural landmarks under conditions where the captured images have sufficient texture. To cope with the fundamental limits of monocular visual odometry (scale ambiguity and rotation-translation ambiguity), vehicle dynamics and airspeed measurements are incorporated in a Kalman filter framework. More robust estimation is provided from multiple rails of the SFM which are traced in an interweaving fashion. Sturdy input to the SFM is enabled by optical flow computation which is tightly coupled with the IMU. Predictive warping parameters and a high-order motion model enhance the accuracy and life span of KLT feature tracking. We also employ vision-based horizon detection as an absolute attitude sensor which is useful for low-level control of a UAV.

The performance of the proposed method is evaluated in what we call an air-slalom task, where the UAV is expected to pass through multiple gates in the air in a row. It will demonstrate how a fixed-wing UAV confronts its limited agility, which is inferior to other hovercraft types in typical urban operations. To efficiently find a feasible obstacle-free path to a goal, we propose a 3D Dubins heuristic for optimal cost to a goal and use a set of lateral and longitudinal motion primitives interconnecting at trim states in order to reduce the dimension of configuration space. We first demonstrate our visual navigation in our UAV simulator, which can be switched between live and synthetic modes, each including wireless data transmission to a ground station.
[full PDF]

Thesis committee:
Takeo Kanade, Co-chair 
James Kuffner, Co-chair 
Sanjiv Singh 
Omead Amidi 
Randy Beard, Brigham Young University

Lab Meeting 09/23, 2009 (Kuo-Huei): Detecting Unusual Activity in Video (CVPR 2004)

Title: Detecting Unusual Activity in Video

Authors: Hua Zhong, Jianbo Shi and Mirko Visontai

Abstract:
We present an unsupervised technique for detecting unusual activity in a large video set using many simple features. No complex activity models and no supervised feature selections are used. We divide the video into equal length segments and classify the extracted features into prototypes, from which a prototype–segment co-occurrence matrix is computed. Motivated by a similar problem in document keyword analysis, we seek a correspondence relationship between prototypes and video segments which satisfies the transitive closure constraint. We show that an important sub-family of correspondence functions can be reduced to co-embedding prototypes and segments to N-D Euclidean space.We prove that an efficient, globally optimal algorithm exists for the co-embedding problem. Experiments on various real-life videos have validated our approach.

Link

CMU talk: Building Vision Systems for Moving Platforms: Background Subtraction from Freely Moving Cameras

Building Vision Systems for Moving Platforms: Background Subtraction from Freely Moving Cameras

Yaser Sheikh
Assistant Research Professor, Robotics, CMU

September 14, 2009, 2:00pm-2:30pm, NSH 3305

Abstract
Most video analysis systems assume staring cameras that continuously view the same scene from the same point of view. Increasingly, as cameras and computers are becoming smaller and cheaper, freely moving cameras are emerging as a primary platform for computer vision research. Background subtraction algorithms, a mainstay in most computer vision systems, define the background as parts of a scene that are at rest. Traditionally, these algorithms assume a stationary camera, and identify moving objects by detecting areas in a video that change over time. In this talk, I will present ideas to extend the concept of ‘subtracting’ areas at rest to apply to video captured from a freely moving camera. We do not assume that the background is well-approximated by a plane or that the camera center remains stationary during motion. The method operates entirely using 2D image measurements without requiring an explicit 3D reconstruction of the scene.

Speaker Biography
Yaser Sheikh is an Assistant Research Professor at the Robotics Institute and Adjunct Professor at the Department of Mechnical Engineering at Carnegie Mellon University. His research is in understanding dynamic scenes through computer vision, including human activity analysis, dynamic scene reconstruction, mobile camera networks, and nonrigid motion estimation. He obtained his doctoral degree from the University of Central Florida in 2006 and is a recipient of the Hillman award for excellence in computer science research.

Lab Meeting 9 / 16, 2009 (Alan): Towards Total Scene Understanding:Classification, Annotation and Segmentation in an Automatic Framework (CVPR 2009)

Title: Towards Total Scene Understanding:Classification, Annotation and Segmentation in an Automatic Framework (CVPR 2009)

Authors: Li-Jia Li, Richard Socher, Li Fei-Fei

Abstract:

Given an image, we propose a hierarchical generative model that classifies the overall scene, recognizes and segments each object component, as well as annotates the image with a list of tags. To our knowledge, this is the first model that performs all three tasks in one coherent framework. For instance, a scene of a ‘polo game’ consists of several visual objects such as ‘human’, ‘horse’, ‘grass’, etc. In addition, it can be further annotated with a list of more abstract (e.g. ‘dusk’) or visually less salient (e.g. ‘saddle’) tags. Our generative model jointly explains images through a visual model and a textual model. Visually relevant objects are represented by regions and patches, while visually irrelevant textual annotations are influenced directly by the overall scene class. We propose a fully automatic learning framework that is able to learn robust scene models from noisy web data such as images and user tags from Flickr.com. We demonstrate the effectiveness of our framework by automatically classifying, annotating and segmenting images from eight classes depicting sport scenes. In all three tasks, our model significantly outperforms stateof-the-art algorithms.

Link: local copy

Paper: Regression-Based Online Situation Recognition for Vehicular Traffic Scenarios

By D. Meyer-Delius, J. Sturm, W. Burgard.

In Proc. of the International Conference on Intelligent Robot Systems (IROS'09), St. Louis, USA, 2009.

Abstract—In this paper, we present an approach for learning generalized models for traffic situations. We formulate the problem using a dynamic Bayesian network (DBN) from which we learn the characteristic dynamics of a situation from labeled trajectories using kernel regression. For a new and unlabeled trajectory, we can then infer the corresponding situation by evaluating the data likelihood for the individual situation models. In experiments carried out on laser range data gathered on a car in real traffic and in simulation, we show that we can robustly recognize different traffic situations even from trajectories corresponding to partial situation instances.

[Full PDF]

CMU PhD Oral: Learning in Modular Systems

CMU PhD Thesis Defense:
David M. Bradley

Title: Learning in Modular Systems

Abstract: Complex robotics systems are often built as a system of modules, where each module solves a separate data processing task to produce the complex overall behavior that is required of the robot. For instance, the perception system for autonomous off-road navigation discussed in this thesis uses a terrain classification module, a ground-plane estimation module, and a path-planning module among others. Splitting a complex task into a series of sub-problems allows human designers to engineer solutions for each sub-problem independently, and devise efficient specialized algorithms to solve them. However, modular design can also create problems for applying learning algorithms. Ideally, learning should find parameters for each module that optimize the performance of the overall system. This requires obtaining ``local'' information for each module about how changing the parameters of that module will impact the output of the system.

Previous work in modular learning showed that if the modules of system were differentiable, gradient descent could be used to provide this local information in “shallow” systems containing with two or three modules between input and output. However, except for convolutional neural networks, this procedure was rarely successful in “deep” systems of more than three modules. Many robotics applications added an additional complication by employing a planning algorithm to produce their output. This makes it hard to define a “loss” function to judge how well the system is performing, or compute a gradient with respect to previous modules in the system.

Recent advances in learning deep neural networks suggest that learning in deep systems can be successful if data-dependent regularization is first used to provide relevant local information to the modules of the system, and the modules are then jointly optimized by gradient descent. Concurrently, research in imitation learning has offered effective new ways of defining loss functions for the output of planning modules.

This thesis combines these lines of research to develop new tools for learning in modular systems. As data-dependent regularization has been shown to be critical to success in deep modular systems, several significant contributions are provided in this area. A novel, differentiable formulation of sparse coding is presented and shown to be a powerful semi-supervised learning algorithm. Sparse coding has traditionally used non-convex optimization methods, and an alternative, convex formulation is developed with a deterministic optimization procedure. Theoretical contributions developed for this convex formulation also enable an efficient, online multi-task learning algorithm. Results in domain adaptation provide further regularization options. To allow joint optimization of systems that employ planning modules, this thesis leverages loss functions developed in recent imitation learning research, and develops techniques for improving all modules of the system with subgradient descent. Finally, this thesis has also made significant contributions to mobile robot perception for navigation, providing terrain classification techniques that been incorporated into fielded industrial and government systems. [Full PDF]


Thesis Committee Members:
James A. Bagnell, Chair
Martial Hebert
Fernando De la Torre
Yoshua Bengio, University of Montreal

CMU Computers and Thought Award Lecture:

How Optimized Environmental Sensing Helps Address Information Overload on the Web

Carlos Guestrin
Finmeccanica Associate Professor
Machine Learning and Computer Science Departments
Carnegie Mellon University


In this talk, we tackle a fundamental problem that arises when using sensors to monitor the ecological condition of rivers and lakes, the network of pipes that bring water to our taps, or the activities of an elderly individual when sitting on a chair: Where should we place the sensors in order to make effective and robust predictions? Such sensing problems are typically NP-hard, and in the past, heuristics without theoretical guarantees about the solution quality have often been used. In this talk, we present algorithms which efficiently find provably near-optimal solutions to large, complex sensing problems. Our algorithms are based on the key insight that many important sensing problems exhibit submodularity, an intuitive diminishing returns property: Adding a sensor helps more the fewer sensors we have placed so far. In addition to identifying most informative locations for placing sensors, our algorithms can handle settings, where sensor nodes need to be able to reliably communicate over lossy links, where mobile robots are used for collecting data or where solutions need to be robust against adversaries and sensor failures. We present results applying our algorithms to several real-world sensing tasks, including environmental monitoring using robotic sensors, activity recognition using a built sensing chair, and a sensor placement competition. We conclude with drawing an interesting connection between sensor placement for water monitoring and addressing the challenges of information overload on the web. As examples of this connection, we address the problem of selecting blogs to read in order to learn about the biggest stories discussed on the web, and personalizing content to turn down the noise in the blogosphere.

Bio: Carlos Guestrin is the Finmeccanica Associate Professor in the Machine Learning and in the Computer Science Departments at Carnegie Mellon University. Previously, he was a senior researcher at the Intel Research Lab in Berkeley. Carlos received his PhD in Computer Science from Stanford University and a Mechatronics Engineer degree from the University of Sao Paulo, Brazil. Carlos' work received awards at a number of conferences and journals. He is also a recipient of the ONR Young Investigator Award, the NSF Career Award, the Alfred P. Sloan Fellowship, and the IBM Faculty Fellowship. He was named one of the 2008 `Brilliant 10' by Popular Science Magazine, received the IJCAI Computers and Thought Award, and the Presidential Early Career Award for Scientists and Engineers (PECASE). Carlos is currently a member of the Information Sciences and Technology (ISAT) advisory group for DARPA.

Lab Meeting August 31, 2009 (swem): Optical moving target detection with 3-D matched filtering

Title: Optical moving target detection with 3-D matched filtering

Author: Reed, Irving S.; Gagliardi, Robert M.; Stotts, Larry B.

IEEE Transactions on Aerospace and Electronic Systems (ISSN 0018-9251), vol. 24, July 1988, p. 327-336.

Abstract:

Three-dimensional (3-D) matched filtering has been suggested as a powerful processing technique for detecting weak, moving optical targets immersed in a background noise field. The procedure requires the processing of entire sequences of frames of optical scenes containing the moving targets. The 3-D processor must be properly matched to the target signature and its velocity vector, but will simultaneously detect all targets to which it is matched. The results of a study to evaluate the 3-D processor are presented. Simulation results are reported which show the ability of the processor to detect targets well below the background level. These results demonstrate the capability and robustness of the processor, and show that the algorithms, although somewhat complicated, can be implemented readily. Some effects on the number of frames processed, target flight scenarios, and velocity and signature mismatch are also presented. The ability to detect multiple targets is demonstrated.

link

Lab Meeting August 31, 2009 (Jim): Maximum Entropy Inverse Reinforcement Learning

I will try to present this paper instead of the previous one.

Title: Maximum Entropy Inverse Reinforcement Learning
B. D. Ziebart, A. Maas, J. A. Bagnell, and A. K. Dey.
AAAI Conference on Artificial Intelligence (AAAI 2008)

Abstract:
In this work, we develop a probabilistic approach based on the principle of maximum entropy. Our approach providesa well-defined, globally normalized distribution over decision sequences, while providing the same performance guarantees as existing methods.

link

Lab Meeting August 31, 2009 (Jim): Acquisition of probabilistic behavior decision model based on the interactive teaching method

T. Inamura, M. Inaba, and H. Inoue
Acquisition of probabilistic behavior decision model based on the interactive teaching method
In Proceedings of the Ninth International Conference on Advanced Robotics(ICAR'99), 1999

Abstract:
In this paper, we propose a novel method for mobile robots to acquire new autonomous behaviors gradually based on interaction between human and robots. In this method, behavior decision models are constructed using statistical process for experiences of interaction and teaching, and the robot expresses sureness of its own decision using stochastic reasoning.
The robot not only decides behavior using the sureness, but also makes suggestions and questions for the user using the sureness.

Link

NTU talk: The Confluence of Sparse Representation and Computer Vision

Title: The Confluence of Sparse Representation and Computer Vision
Speaker: Professor Yi Ma, ECE Department, UIUC and Microsoft Research Asia
Time: 2:20pm, Aug 28 (Fri), 2009
Place: Room 101, CSIE building

Abstract:

In the past few years, sparse representation and compressive sensing have arisen as a very powerful and popular framework for signal and image processing. It has armed people with new mathematical principles and computational tools that can effectively and efficiently harness sparse, low-dimensional structures of high-dimensional data such as images and videos. In this talk, we contend that the same principles and tools are equally important for analyzing the meaning and semantics of images and help solve many outstanding problems in computer vision.

As an example, we will focus on the recent success of sparse representation in human face recognition. On one hand, tools from sparse representation such as L1-minimization have seen great empirical success in enhancing the robustness of face recognition with occlusion, illumination change, and registration error, leading to striking recognition performance far exceeding human expectation or capability. On the other hand, the peculiar structures of face images have led to new mathematical discovery of remarkable properties of L1 minimization that far exceed the existing sparse representation theory.

We will also illustrate with many other examples in computer vision the importance of sparsity as a guiding principle for extracting and harnessing the structures of high-dimensional visual data. In return, we will see that overwhelming empirical evidences from those examples suggest that an even richer set of new mathematical results can be developed if we systematically extend the theory of sparse representation to clustering or classification of high-dimensional visual data. The confluence of sparse representation and computer vision is leading us to a brand new mathematical foundation for high-dimensional pattern analysis and recognition.

This is joint work with my former PhD students John Wright, Allen Yang, and Shankar Rao.


Short Biography:

Yi Ma is an associate professor at the Electrical & Computer Engineering Department of the University of Illinois at Urbana-Champaign. He is currently on leave as research manager of the Visual Computing group at Microsoft Research Asia in Beijing. His research interests include computer vision, image processing, and systems theory. Yi Ma received two Bachelors’ degree in Automation and Applied Mathematics from Tsinghua University (Beijing, China) in 1995, a Master of Science degree in EECS in 1997, a Master of Arts degree in Mathematics in 2000, and a PhD degree in EECS in 2000, all from the University of California at Berkeley. Yi Ma received the David Marr Best Paper Prize at the International Conference on Computer Vision 1999 and the Longuet-Higgins Best Paper Prize at the European Conference on Computer Vision 2004. He also received the CAREER Award from the National Science Foundation in 2004 and the Young Investigator Award from the Office of Naval Research in 2005. He is an associate editor of IEEE Transactions on Pattern Analysis and Machine Intelligence. He is a senior member of IEEE and a member of ACM, SIAM, and ASEE.

Beyond Asimov: The Three Laws of Responsible Robotics

by Robin R. Murphy and David D. Woods

IEEE Intelligent Systems, July/August 2009, pp. 14–20

Since their codification in 1947 in the collection of short stories I, Robot, Isaac Asimov’s three laws of robotics have been a staple of science fiction. Most of the stories assumed that the robot had complex perception and reasoning skills equivalent to a child and that robots were subservient to humans. Although the laws were simple and few, the stories attempted to demonstrate just how difficult they were to apply in various real-world situations. In most situations, although the robots usually behaved "logically," they often failed to do the "right" thing, typically because the particular context of application required subtle adjustments of judgment on the part of the robot (for example, determining which law took priority in a given situation, or what constituted helpful or harmful behavior).

[The full article]

Lab Meeting August 24, 2009(Chung-Han) : Monitoring an intersection using a network of laser scanners

Monitoring an intersection using a network of laser scanners

Huijing Zhao; Jinshi Cui; Hongbin Zha; Katabira, K.; Xiaowei Shao; Shibasaki, R.

Intelligent Transportation Systems, 2008. ITSC 2008. 11th International IEEE Conference on12-15 Oct. 2008 Page(s):428 - 433

Abstract : In this research, a novel system for monitoring an intersection using a network of single-row laser range scanners (subsequently abbreviated as "laser scanner") is proposed. Laser scanners are set on the road side to profile an intersection horizontally from different viewpoints. This is done so that cross sections of the intersection are captured at a high scanning rate (e.g., 37 Hz) and to contain the contour points of the moving objects entering the intersection. Different laser scanners data are integrated into a common spatial-temporal coordinate system and processed. Thus, the moving objects inside the intersection are detected and tracked to estimate their state parameters, such as: location, speed, and direction at each time instance. An experiment was conducted in central Beijing, where six laser scanners were used to cover a three-way intersection. A digital copy of the dynamic intersection was measured, and, through data processing, a large quantity of physical dimension traffic data was obtained.

[link]

Lab Meeting August 24, 2009(Jimmy): One-Shot Learning of Object Categories

Title: One-Shot Learning of Object Categories

In: IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 28, No. 4, April 2006

Li Fei-Fei, Rob Fergus, and Pietro Perona

Abstract: Learning visual models of object categories notoriously requires hundreds or thousands of training examples. We show that it is possible to learn much information about a category from just one, or a handful, of images. The key insight is that, rather than learning from scratch, one can take advantage of knowledge coming from previously learned categories, no matter how different these categories might be. We explore a Bayesian implementation of this idea. Object categories are represented by probabilistic models. Prior knowledge is represented as a probability density function on the parameters of these models. The posterior model for an object category is obtained by updating the prior in the light of one or more observations. We test a simple implementation of our algorithm on a database of 101 diverse object categories. We compare category models learned by an implementation of our Bayesian approach to models learned from by Maximum Likelihood (ML) and Maximum A Posteriori (MAP) methods. We find that on a database of more than 100 categories, the Bayesian approach produces informative models when the number of training examples is too small for other methods to operate successfully.

[Link]

Lab Meeting August 17, 2009 (Any): RANSAC-based DARCES

RANSAC-Based DARCES: A New Approach to Fast Automatic Registration of Partially Overlapping Range Images

IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), Vol. 21, No. 11, November 1999

Chu-Song Chen, Yi-Ping Hung, and Jen-Bo Cheng

Abstract: In this paper, we propose a new method, the RANSAC-based DARCES method, which can solve the partially overlapping 3D registration problem without any initial estimation. For the noiseless case, the basic algorithm of our method can guarantee that the solution it finds is the true one, and its time complexity can be shown to be relatively low. An extra characteristic is that our method can be used even for the case that there are no local features in the 3D data sets.

Link

Lab Meeting August 17, 2009(Kuen-Han): Moving Obstacle Detection in Highly Dynamic Scenes(ICRA09)

Title:Moving Obstacle Detection in Highly Dynamic Scenes(ICRA09)
Authors:A. Ess, B. Leibe, K. Schindler, and L. van Gool.

Abstract

We address the problem of vision-based multipersontracking in busy pedestrian zones using a stereo rigmounted on a mobile platform. Specifically, we are interestedin the application of such a system for supporting pathplanning algorithms in the avoidance of dynamic obstacles.
The complexity of the problem calls for an integrated solution, whichextracts as much visual information as possible and combinesit through cognitive feedback.

We propose such an approach,which jointly estimates camera position, stereo depth, objectdetections, and trajectories based only on visual information.The interplay between these components is represented in agraphical model. For each frame, we first estimate the groundsurface together with a set of object detections. Based onthese results, we then address object interactions and estimatetrajectories. Finally, we employ the tracking results to predictfuture motion for dynamic objects and fuse this informationwith a static occupancy map estimated from dense stereo.

The approach is experimentally evaluated on several longand challenging video sequences from busy inner-city locationsrecorded with different mobile setups. The results show thatthe proposed integration makes stable tracking and motionprediction possible, and thereby enables path planning incomplex and highly dynamic scenes.

link

webpage

NewScientist: Why humans can't navigate out of a paper bag

12 August 2009 by Chris Berdik
Magazine issue 2721.
[Full Article]

This is an interesting article mentioning that human beings do not perform metric-SLAM and may perform topological SLAM poorly. Take a look.

-Bob

IJCAI 2009 Talk: From Low-level Sensors to High-level Intelligence: Activity Recognition Links the Knowledge Food Chain

Title: From Low-level Sensors to High-level Intelligence: Activity Recognition Links the Knowledge Food Chain (IJCAI 2009)

Author: Qiang Yang, The Hong Kong University of Science and Technology

Description
Sensors provide computer systems with a window to the outside world. Activity recognition "sees" what is in the window to predict the locations, trajectories, actions, goals and plans of humans and objects. Building an activity recognition system requires a full range of interaction from statistical inference on lower level sensor data to symbolic AI at higher levels, where prediction results and acquired knowledge are passed up each level to form a knowledge food chain. In this talk, I will give an overview of activity recognition and explore its relation to other fields, including planning and knowledge acquisition, machine learning and Web search. I will also describe its applications in assistive technologies, security monitoring and mobile commerce.

Link

Lab Meeting August 10, 2009(Nicole):Learning Sound Location from a Single Microphone (ICRA 2009)

Title:Learning Sound Location from a Single Microphone (ICRA 2009)

Authors:AshutoshSaxena and AndrewY. Ng

Abstract:

We consider the problem of estimating the incident angle of a sound, using only a single microphone. The ability to perform monaural (single-ear) localization isimportant to many animals; indeed, monaural cues are also the primary method by which humans decide if a sound comes from the front or back, as well as estimate its elevation. Such monaural localization is made possible by the structure of the pinna (outer ear), which modifies sound in a way that is dependent on its incident angle. In this paper, we propose a machine learning approach to monaural localization, using only a single microphone and an “artificial pinna” (that distorts sound in a direction-dependent way). Our approach models the typical distribution of natural and artificial sounds, as well as the direction-dependent changes to sounds induced by the pinna. Our experimental results also show that the algorithm is able to fairly accurately localize a wide range of sounds, such as human speech, dog barking, waterfall, thunder, and so on. In contrast to microphone arrays, this approach also offers the potential of significantly more compact, as well as lower cost and power, devices for sounds localization.

[link]

Paper: Modeling Groups of Plausible Virtual Pedestrians

by Christopher Peters and Cathy Ennis
IEEE Computer Graphics and Applications, July/August 2009, pp. 54–63

Crowd simulation is enjoying considerable success in a number of applied domains, most notably in evacuation scenarios in which simulated crowd behaviors can help improve the safety of interior building designs. However, not all applications involving virtual populace have the overarching goal of realistic simulation. In many cases, it's necessary only that viewers perceive the crowd as realistic. In many of the latest movies or video games involving large numbers of virtual actors, liberties can be taken in displaying those far away or otherwise obscured from the eye, if this doesn't noticeably diminish the viewing experience. For example, such simulations can reduce the level of detail or forgo collision avoidance calculations to allow simulation of a larger crowd or enhanced behaviors for individuals deemed most likely to occupy viewers' attention. (Full PDF)

FRC Seminar Special Time:Information Sharing in Large Heterogeneous Teams, August 13, 2009

FRC Seminar

August 13, 2009 12pm

Speaker:

Prasanna Velagapudi
PhD Student
Robotics Institute
Carnegie Mellon University

Abstract:

In large, collaborative, heterogeneous teams, team members often collect information that is useful to other members of the team. However, in many real domains, it is impossible to completely share all of this information due to network and processing constraints. Recognizing the utility of such information and delivering it efficiently across a team has been the focus of much research, with proposed approaches ranging from flooding to complex filters and matchmakers. Interestingly, random forwarding of information has been found to be a surprisingly effective information sharing approach in some domains. In this talk, we investigate some recent results supporting this phenomenon in detail and show that in certain systems, random forwarding of information performs almost half as well as a globally optimal approach. From this, we demonstrate a statistical modeling approach designed to estimate information sharing performance in real domains. Finally, we will discuss ongoing work and possible applications of these models in enabling heterogeneous teams to be scaled into the 100s and 1000s.

Speaker Bio:

Prasanna is currently pursuing a PhD at the Robotics Institute, and is co-advised by Katia Sycara and Paul Scerri. His research focuses on information sharing in large heterogeneous teams and large-scale human-robot interaction. Previously, he worked as an electrical engineer at RedZone Robotics, developing on power systems and embedded computing for submersible, subterranean mapping and mobility applications. Prasanna holds a B.S. in Electrical and Computer Engineering and Computer Science from Carnegie Mellon University.

MIT CSAIL Thesis Defense: Visual Sense Disambiguation: A Multimodal Approach

Speaker: Kate Saenko, MIT CSAIL

Date: Friday, August 7 2009

Time: 2:00PM to 3:00PM

Refreshments: 1:45PM

Location: Star conference room

Contact: Kate Saenko, (617) 669-9093, saenko@mit.edu

If a picture is worth a thousand words, can a thousand words be worth a training image? Most successful object recognition algorithms require manually annotated images of objects to be collected for training. The amount of human effort required to collect training data has limited most approaches to the several hundred object categories available in the labeled datasets. While human-annotated image data is scarce, additional sources of information can be used as weak labels, reducing the need for human supervision. In this thesis, we use three types of information to learn models of object categories: speech, text and dictionaries. We demonstrate that our use of non-traditional information sources facilitates automatic acquisition of visual object models for arbitrary words without requiring any labeled image examples.

Spoken object references occur in many scenarios: interaction with an assistant robot, voice-tagging of photos, etc. Existing reference resolution methods are unimodal, relying either only on image features, or only on speech recognition. We propose a method that uses both the image of the object and the speech segment referring to it to disambiguate the underlying object label. We show that even noisy speech input helps visual recognition, and vice versa. We also explore two sources of linguistic sense information: the words surrounding images on web pages, and dictionary entries for nouns that refer to objects. Keywords that index images on the web have been used as weak object labels, but these tend to produce noisy datasets with many unrelated images. We use unlabeled text, dictionary definitions, and semantic relations between concepts to learn a refined model of image sense. Our model can work with as little supervision as a single English word. We apply this model to a dataset of web images indexed by polysemous keywords, and show that it improves both retrieval of specific senses, and the resulting object classifiers.

Link

Lab Meeting August 10, 2009(Gary):In Between 3D Active Appearance Models and 3D Morphable Models

Title:In Between 3D Active Appearance Models and 3D Morphable Models
Authors: Jingu Heo, Marios Savvides

Abstract:
In this paper we propose a novel method of generating
3D Morphable Models (3DMMs) from 2D images. We
develop algorithms of 3D face reconstruction from a
sparse set of points acquired from 2D images. In order
to establish correspondence between images precisely, we
Combined Active Shape Models (ASMs) and Active Appearance
Models (AAMs)(CASAAMs) in an intelligent way,
showing improved performance on pixel-level accuracy
and generalization to unseen faces. The CASAAMs are
applied to the images of different views of the same person
to extract facial shapes across pose. These 2D shapes are
combined for reconstructing a sparse 3D model. The point
density of the model is increased by the Loop subdivision
method, which generates new vertices by a weighted sum
of the existing vertices. Then, the depth of the dense 3D
model is modified with an average 3D depth-map in order
to preserve facial structure more realistically. Finally, all
249 3D models with expression changes are combined
to generate a 3DMM for a compact representation. The
first session of the Multi-PIE database, consisting of 249
persons with expression and illumination changes, is used
for the modeling. Unlike typical 3DMMs, our model can
generate 3D human faces more realistically and efficiently
(2-3 seconds on P4 machine) under diverse illumination
conditions.


link

Lab Meeting August 3rd, 2009 (Jeff): Angular problem after loop closing

I will introduce the angular problem after loop closing and discuss with you.

And try to propose some methods to slove this problem.

MIT Talk: Biomedical Imaging and Analysis Seminar

Biomedical Imaging and Analysis Seminar

Speaker: Arno Klein , Columbia University

Date: Friday, July 31 2009

Time: 2:00PM to 3:00PM

Location: 32-D507

Host: Thomas Yeo, CSAIL

Contact: Thomas Yeo, 617-253-4143, ythomas@csail.mit.edu

Establishing correspondences across brains for the purposes of comparison and group analysis is almost universally done by registering images to one another either directly or via a template. However, there are many registration algorithms to choose from. The first part of this talk will give an overview of a recent evaluation study comparing fully automated nonlinear deformation methods applied to brain image registration (Klein et al. 2009). This study was restricted to volume-based methods, and an ongoing extension of this study is the first known to the authors that directly compares some of the most accurate of these methods with surface-based registration methods, as well as the first study to compare registrations of whole-head and de-skulled brain images. More than 6,000 registrations between 40 manually labeled brain images have been performed so far by the volume-based algorithms ART and SyN and the surface-based algorithms FreeSurfer and Spherical Demons. We used permutation tests and indifference-zone ranking to compare the overlap performance for eight scenarios: ART and SyN on brain images with and without skulls, SyN, FreeSurfer, and Spherical Demons via custom templates, and FreeSurfer via its default atlas.

Link

Phd oral exam : Large Scale Scene Matching for Graphics and Vision

July 29, 2009
James H. Hays
11:00 AM, 7220 Wean Hall
Thesis Oral
Title: Large Scale Scene Matching for Graphics and Vision

Abstract:
Our visual experience is extraordinarily varied and complex. Thediversity of the visual world makes it difficult for computervision to understand images and for computer graphics tosynthesize visual content. But for all its richness, it turns outthat the space of "scenes" might not be astronomically large.With access to imagery on an Internet scale, regularities start toemerge - for most images, there exist numerous examples ofsemantically and structurally similar scenes. Is it possible tosample the space of scenes so densely that one can use similarscenes to "brute force" otherwise difficult image understandingand manipulation tasks? This thesis is focused on exploiting andrefining large scale scene matching to short circuit thetypical computer vision and graphics pipelines for imageunderstanding and manipulation.

First, in "Scene Completion" we patch up holes in images bycopying content from matching scenes. We find scenes so similarthat the manipulations are undetectable to naive viewers and wequantify our success rate with a perceptual study. Second, in"im2gps" we estimate geographic properties and globalgeolocation for photos using scene matching with a database of 6million geo-tagged Internet images. We geolocate sequences ofphotos four times as accurately as the single image case bymodelling the global spatiotemporal statistics of photographers.We introduce a range of features for scene matching and use them,together with lazy SVM learning, to dramatically improve scenematching -- doubling the performance of single image geolocationover our baseline method. Third, we study human photo geolocationto gain insights into the geolocation problem, our algorithms, andhuman scene understanding. This study shows that our algorithmssignificantly exceed human geolocation performance. Finally, weuse our geography estimates, as well as Internet text annotations,to provide context for deeper image understanding, such as objectdetection.

Thesis Committee:Alexei A. Efros, ChairMartial HebertJessica K. HodginsTakeo KanadeRichard Szeliski, Microsoft Research

Lab Meeting August 3, 2009(Shao-Chen): Time-bounded Lattice for Efficient Planning in Dynamic Environments(ICRA09)

Title:Time-bounded Lattice for Efficient Planning in Dynamic Environments(ICRA09)

Authors:Aleksandr Kushleyev, Maxim Likhachev

Abstract

For vehicles navigating initially unknown cluttered environments, current state-of-the-art planning algorithms are able to plan and re-plan dynamically-feasible paths efficiently and robustly. It is still a challenge, however, to deal well with the surroundings that are both cluttered and highly dynamic. Planning under these conditions is more difficult for two reasons. First, tracking and predicting the trajectories of moving objects (i.e., cars, humans) is very noisy. Second, the planning process is computationally more expensive because of the increased dimensionality of the state-space, with time as an additional variable. Moreover, re-planning needs to be invoked more often since the trajectories of moving obstacles need to be constantly re-estimated.

In this paper, we develop a path planning algorithm that addresses these challenges. First, we choose a representation of dynamic obstacles that efficiently models their predicted trajectories and the uncertainty associated with the predictions. Second, to provide real-time guarantees on the performance of planning with dynamic obstacles, we propose to utilize a novel data structure for planning - a time-bounded lattice - that merges together short-term planning in time with longterm planning without time. We demonstrate the effectiveness of the approach in both simulations with up to 30 dynamic obstacles and on real robots.

link

Lab Meeting July 13, 2009 (Alan): An Embodied Cognition Approach to Mindreading Skills for Socially Intelligent Robots (IJRR 2009)

Title: An Embodied Cognition Approach to Mindreading Skills for Socially Intelligent Robots (IJRR 2009)
Authors: Cynthia Breazeal, Jesse Gray, Matt Berlin

Abstract
Future applications for personal robots motivate research into developing robots that are intelligent in their interactions with people. Toward this goal, in this paper we present an integrated socio-cognitive architecture to endow an anthropomorphic robot with the ability to infer mental states such as beliefs, intents, and desires from the observable behavior of its human partner. The design of our architecture is informed by recent findings from neuroscience and embodies cognition that reveals how living systems leverage their physical and cognitive embodiment through simulation-theoretic mechanisms to infer the mental states of others. We assess the robot’s mindreading skills on a suite of benchmark tasks where the robot interacts with a human partner in a cooperative scenario and a learning scenario. In addition, we have conducted human subjects experiments using the same task scenarios to assess human performance on these tasks and to compare the robot’s performance with that of people. In the process, our human subject studies also reveal some interesting insights into human behavior.

Link

Lab Meeting 7/13, 2009(Casey): A Closed-Form Solution to Non-Rigid Shape and Motion Recovery

Title: A Closed-Form Solution to Non-Rigid Shape and Motion Recovery (IJCV 2006)

Authors: J. Xiao, J.Chai, and T.Kanade

Abstract:

Abstract. Recovery ofthree dimensional (3D) shape and motion ofnon-static scenes from a monocular video sequence is important for applications like robot navigation and human computer interaction. Ifev ery point in the scene randomly moves, it is impossible to recover the non-rigid shapes. In practice, many non-rigid objects, e.g. the human face under various expressions, deform with certain structures. Their shapes can be regarded as a weighted combination of certain shape bases. Shape and motion recovery under such situations has attracted much interest. Previous work on this problem [6, 4, 14] utilized only orthonormality constraints on the camera rotations (rotation constraints). This paper proves that using only the rotation constraints results in ambiguous and invalid solutions. The ambiguity arises from the fact that the shape bases are not unique. An arbitrary linear transformation of the bases produces another set ofeligible bases. To eliminate the ambiguity, we propose a set of novel constraints, basis constraints, which uniquely determine the shape bases. We prove that, under the weak-perspective projection model, enforcing both the basis and the rotation constraints leads to a closed-form solution to the problem of non-rigid shape and motion recovery. The accuracy and robustness ofour closed-form solution is evaluated quantitatively on synthetic data and qualitatively on real video sequences.

[Link]

Lab Meeting 7/13, 2009(fish60): Autonomous Robot Navigation in Outdoor Cluttered Pedestrian Walkways

Autonomous Robot Navigation in Outdoor Cluttered Pedestrian Walkways
Yoichi Morales, Alexander Carballo, Eijiro Takeuchi, Atsushi Aburadani, and Takashi Tsubouchi
Journal of Field Robotics 26(8), 609–635 (2009)

Abstract:
This paper describes an implementation of a mobile robot system for autonomous navigationin outdoor concurred walkways. The task was to navigate through nonmodified pedestrian paths with people and bicycles passing by. The proposed approach proved to be robust for outdoor navigationin cluttered and crowded walkways, first on campus paths and then running the challenge course multiple times between trials and the challenge final. The paper reports experimental results and overall performance of the system. Finally the lessons learned are discussed. The main contribution of this work is the report of a system integration approach for autonomous outdoor navigation and its evaluation.

Link

News: Language may be key to theory of mind

Language may be key to theory of mind

• 15:55 23 June 2009 by Anil Ananthaswamy
• For similar stories, visit the The Human Brain Topic Guide

How blind and deaf people approach a cognitive test regarded as a milestone in human development has provided clues to how we deduce what others are thinking.

Understanding another person's perspective, and realising that it can differ from our own, is known as theory of mind. It underpins empathy, communication and the ability to deceive – all of which we take for granted. Although our theory of mind is more developed than it is in other animals, we don't acquire it until around age four, and how it develops is a mystery.

See the full article.

Lab Meeting June 21th, 2009 (swem): Improved Inverse-Depth Parameterization for Monocular Simultaneous Localization and Mapping

Title: Improved Inverse-Depth Parameterization for Monocular
Simultaneous Localization and Mapping
In: ICRA2009

Author: E. Imre, M.-O. Berger, N. Noury

Abstract:

Inverse-depth parameterization can successfully
deal with the feature initialization problem in monocular
simultaneous localization and mapping applications. However,
it is redundant, and when multiple landmarks are initialized
from the same image, it fails to enforce the “common origin”
constraint. The authors propose two new variants that
addresses both of these issues. The experimental results indicate
that the proposed approach achieves a better performance at a
lower computational cost.

[ Link ]

Intelligence Seminar: Action Perception, June 23, 2009

Intelligence Seminar

June 23, 2009 (note special place)
3:30 pm
NSH 1507
Host: Jaime Carbonell
For meetings, contact Michelle Pagnani (pagnani@cs.cmu.edu).

Action Perception

Robert Thibadeau
Seagate Research

Abstract:
The human perception of actions has barely been studied, but this study of action perception promises to provide a wealth of interesting hypotheses regarding cognitive processing. Action perception is distinct from motion perception in that the direct perception of causation is central to the percept. Among the interesting hypotheses is that it can be hypothesized that what we know as thought and reasoning is where we perceive and plan actions. Another hypothesis is that what we know as logic and mathematics derives from our direct perceptions of causation in the actions we perceive and think about.

I will present a study that attempts to estimate the scale of computation needed to implement a system for visually perceiving meaningful actions and non-trivially producing an English narration of what is being visually perceived, as well as answering questions about what is visually perceived. The scale of the computation for learning could easily reach exaflops over distributed datasets (HADOOP or MapReduce style).

This study is partly based on my work (Thibadeau, 1986), and Doug Rohde's 2002 dissertation (http://tedlab.mit.edu:16080/~dr/Thesis/), as well as Simon and Rescher (1966 see summary below). The study includes an explicit proposal for extending Rohde's work to multimodal, multisensory, processing.

(Simon and Rescher 1966 From Wikipedia, Causality)
Derivation theories

The Nobel Prize holder Herbert Simon and Philosopher Nicholas Rescher[20] claim that the asymmetry of the causal relation is unrelated to the asymmetry of any mode of implication that
contraposes. Rather, a causal relation is not a relation between values of variables, but a function of one variable (the cause) on to another (the effect). So, given a system of equations, and a set of
variables appearing in these equations, we can introduce an asymmetricrelation among individual equations and variables that corresponds perfectly to our commonsense notion of a causal ordering. The system of equations must have certain properties, most importantly, if some values are chosen arbitrarily, the remaining values will be determined uniquely through a path of serial discovery that is perfectly causal. They postulate the inherent serialization of such a system of equations may correctly capture causation in all empirical fields, including physics and economics.

RSS 2009 paper: Non-parametric Learning To Aid Path Planning Over Slopes

Title: Non-parametric Learning To Aid Path Planning Over Slopes (RSS 2009)

Sisir Karumanchi, Thomas Allen, Tim Bailey and Steve Scheding
ARC Centre of Excellence For Autonomous Systems (CAS),
Australian Centre For Field Robotics (ACFR),
The University of Sydney,
NSW. 2006, Australia.


Abstract—This paper addresses the problem of closing the loop from perception to action selection for unmanned ground vehicles, with a focus on navigating slopes. A new non-parametric learning technique is presented to generate a mobility representation where maximum feasible speed is used as a criterion to classify the world. The inputs to the algorithm are terrain gradients derived from an elevation map and past observations of wheel slip. It is argued that such a representation can aid in path planning with improved selection of vehicle heading and operating velocity in off-road slopes. Results of mobility map generation and its benefits to path planning are shown.

Link

RSS 2009 Paper: Generalized-ICP

Title: Generalized-ICP (RSS 2009)

Authors: Aleksandr Segal, Dirk Haehnel, Sebastian Thrun

Abstract:

In this paper we combine the Iterative Closest Point (’ICP’) and ‘point-to-plane ICP‘ algorithms into a single probabilistic frame work. We then use this framework to model locally planar surface structure from both scans instead of just the ”model” scan as is typically done with point-to-plane. This can be thought of as ‘plane-to-plane’. The new approach is tested with both simulated and real-world data and is shown to outperform both standard ICP and point-to-lane. Furthermore, the new approach is shown to be more robust to incorrect correspondences, and thus makes it easier to tune the maximum match distance parameter. In addition to the demonstrated performance improvement, the proposed framework allows for more expressive probabilistic models to be incorporated into the ICP framework. While maintaining the speed and simplicity of ICP, Generalized-ICP allows the addition of outlier terms, measurement noise, and other probabilistic techniques to increase robustness.

Link

RSS 2009 Paper: Large Scale Graph-based SLAM using Aerial Images as Prior Information

Title: Large Scale Graph-based SLAM using Aerial Images as Prior Information (RSS 2009)

Authors: Rainer K¨ummerle Bastian Steder Christian Dornhege Alexander Kleiner Giorgio Grisetti Wolfram Burgard

Abstract:

To effectively navigate in their environments and accurately reach their target locations, mobile robots require a globally consistent map of the environment. The problem of learning a map with a mobile robot has been intensively studied in the past and is usually referred to as the simultaneous localization and mapping (SLAM) problem. However, existing solutions to the SLAM problem typically rely on loop-closures to obtain global consistency and do not exploit prior information even if it is available. In this paper, we present a novel SLAM approach that achieves global consistency by utilizing publicly accessible aerial photographs as prior information. Our approach inserts correspondences found between three-dimensional laser range scans and the aerial image as constraints into a graph-based formulation of the SLAM problem. We evaluate our algorithm based on large real-world datasets acquired in a mixed in- and outdoor environment by comparing the global accuracy with state-of-the-art SLAM approaches and GPS. The experimental results demonstrate that the maps acquired with our method show increased global consistency.

Link

Lab Meeting June 15th, 2009 (Andi): Progress Report

I will talk about my recent progress in 3D Mapping and Localization using one 2D LIDAR.

Using the Distribution Theory to Simultaneously Calibrate the Sensors of a Mobile Robot

Title: Using the Distribution Theory to Simultaneously Calibrate the Sensors of a Mobile Robot

Author: Agostino Martinelli

Abstract:
This paper introduces a simple and very efficient strategy to extrinsically calibrate a bearing sensor (e.g. a camera) mounted on a mobile robot and simultaneously estimate the parameters describing the systematic error of the robot odometry system. The paper provides two contributions. The first one is the analytical computation to derive the part of the system which
is observable when the robot accomplishes circular trejectories. This computation consists in performing a local decomposition of the system, based on the theory of distributions. In this respect, this paper represents the first application of the distribution theory in the frame-work of mobile robotics. Then, starting from this decomposition, a method to efficiently estimate the parameters describing both the extrinsic bearing sensor calibration and the odometry calibration is derived (second contribution). Simulations and experiments with the robot e-Puck equipped
with encoder sensors and a camera validate the approach.

Link:
RSS05
http://www.roboticsproceedings.org/rss05/p11.pdf
http://hal.inria.fr/docs/00/35/30/79/PDF/RR-6796.pdf

News: Willow Garage: Milestone 2 Start to Finish

Lab Meeting June 8th, 2009(ZhenYu):Vertical Line Matching for Omnidirectional Stereovision Images

Title: Vertical Line Matching for Omnidirectional Stereovision Images (ICRA2009)

Authors: Guillaume Caron and El Mustapha Mouaddib

Abstract: We are investigating the mobile robot indoor localization and environment mapping using an omnidirectional stereovision sensor. It uses four parabolic mirrors and an orthographic camera, giving four images of the same scene. At least, only two mirrors are needed. Using four mirrors gives redundancy. We propose to exploit the images of vertical lines. This paper presents a new method in order to match these lines in the four images. Contrary to existing approaches, we took into account the four sub-images existence in the design of this method, in order to exploit redundancy. This brought an original algorithm combining matching and pose estimation of vertical lines from the 3D environment. Experimental results will be presented to validate this approach.

[Link]

Lab Meeting June 8th, 2009 (Any): CRF-Filters

Paper title: CRF-Filters: Discriminative Particle Filters for Sequential State Estimation

Authors: Benson Limketkai, Dieter Fox and Lin Liao

Appears in: ICRA 2007

Abstract: Particle filters have been applied with great success to various state estimation problems in robotics. However, particle filters often require extensive parameter tweaking in order to work well in practice. This is based on two observations. First, particle filters typically rely on independence assumptions such as “the beams in a laser scan are independent given the robot’s location in a map”. Second, even when the noise parameters of the dynamical system are perfectly known, the sample-based approximation can result in poor filter performance. In this paper we introduce CRF-Filters, a novel variant of particle filtering for sequential state estimation. CRF-Filters are based on conditional random fields, which are discriminative models that can handle arbitrary dependencies between observations. We show how to learn the parameters of CRF-Filters based on labeled training data. Experiments using a robot equipped with a laser range-finder demonstrate that our technique is able to learn parameters of the robot’s motion and sensor models that result in good localization performance, without the need of additional parameter tweaking.

Full text: PDF

Some talks at ICRA 2009.

Talks at three forums (Industrial Forum, Science Forum, and Citizen's Forum) and plenary talks of IEEE ICRA 2009 are posted at http://podcasts.icra2009.org/rss.xml.

ICRA2009 Awards

Best Vision Paper
Moving Obstacle Detection in Highly Dynamic Scenes [Local Copy][Attachment]

Best Automation Paper and Best Student Paper
Design and Calibration of a Microfabricated 6-Axis Force-Torque Sensor for Microrobotic Applications [Local Copy]

Best Conference Paper
Towards a Navigation System for Autonomous Indoor Flying [Local Copy]

Lab Meeting June 1st, 2009 (Jeff): Modeling RFID Signal Strength and Tag Detection for Localization and Mapping

Title: Modeling RFID Signal Strength and Tag Detection for Localization and Mapping

Authors: Dominik Joho, Christian Plagemann and Wolfram Burgard

Abstract:

In recent years, there has been an increasing interest within the robotics community in investigating whether Radio Frequency Identification (RFID) technology can be utilized to solve localization and mapping problems in the context of mobile robots. We present a novel sensor model which can be utilized for localizing RFID tags and for tracking a mobile agent moving through an RFID-equipped environment. The proposed probabilistic sensor model characterizes the received signal strength indication (RSSI) information as well as the tag detection events to achieve a higher modeling accuracy compared to state-of-the-art models which deal with one of these aspects only. We furthermore propose a method that is able to bootstrap such a sensor model in a fully unsupervised fashion. Real-world experiments demonstrate the effectiveness of our approach also in comparison to existing techniques.

Link:
ICRA2009
http://www.informatik.uni-freiburg.de/~joho/publications/joho09icra.html

Lab Meeting May 25, 2009 (fish60):

I may talk about what I read in recent weeks based on the following paper.

Brenna Argall, Sonia Chernova and Manuela Veloso. A Survey of Robot Learning from Demonstration. Robotics and Autonomous Systems. Vol. 57, No. 5, pages 469-483, 2009.

Link

abstract:
We present a comprehensive survey of robot Learning from Demonstration (LfD), a technique that develops policies from example state to action mappings. We introduce the LfD design choices in terms of demonstrator, problem space, policy derivation and performance, and contribute the foundations for a structure in which to categorize LfD research. Specifically, we analyze and categorize the multiple ways in which examples are gathered, ranging from teleoperation to imitation, as well as the various techniques for policy derivation, including matching functions, dynamics models and plans. To conclude we discuss LfD limitations and related promising areas for future research.

MIT CSAIL Technical Report: Scene Classification with a Biologically Inspired Method

Title: Scene Classification with a Biologically Inspired Method

(MIT CSAIL Technical Report)

Abstract:

We present a biologically motivated method for scene image classification. The core of the method is to use shape based image property that is provided by a hierarchical feedforward model of the visual cortex [18]. Edge based and color based image properties are additionally used to improve the accuracy. The method consists of two stages of image analysis. In the first stage, each of three paths of classification uses each image property (i.e. shape, edge or color based features) independently. In the second stage, a single classifier assigns the category of an image based on the probability distributions of the first stage classifier outputs. Experiments show that the method boosts the classification accuracy over the shape based model. We demonstrate that this method achieves a high accuracy comparable to other reported methods on publicly available color image dataset.

[download]

VASC Seminar: A Data-Driven Vision Compiler for Automatic Object Pose Recognition

VASC Seminar
Monday, May 11, 2009
3:30p-4:30p
NSH 1507

A Data-Driven Vision Compiler for Automatic Object Pose Recognition

Rosen Diankov
Carnegie Mellon University, Robotics Institute


Abstract:

This presentation focuses on an object-specific vision system that detects and extracts the precise 6D pose of objects in an image. The system builds a data-driven statistical model of the expected features of an object's surface and combines this with a discrete search method
to extract the pose of all object. The training phase of the vision system can be interpreted as a compiler that automatically analyzes the statistics of how the features are distributed on the object and determines a feature set's stability and discriminable power. This compilation phase requires the precise CAD model of an object along with a training set of real-world images. After compilation, a CAD-independent model of how features relate with respect to theobject's pose and inter-relate with each other is created. These relationships allow both point-based features like SIFT and edge-based features to be used simultaneously when computing the 6D pose of an
object. Using this data-driven model, we employ a discrete randomized search with RANSAC to find the poses of all instances of the object in a novel image.


Bio:

Rosen Diankov graduated from University of California Berkeley in 2006 with Electrical Engineering and Computer Science, and Applied Math degrees. At the moment he is a PhD graduate student at the Robotics Institute at Carnegie Mellon University. Rosen's main research focus is tackling the robotics problem: combining perception, planning, and control into one coherent framework. Up until now he has worked on several vision and planning systems involving autonomous robots in everyday scenarios both in the United States and Japan.

VASC Seminars are sponsored by Tandent Vision Science, Inc.

Lab Meeting May 25, 2009(Chung-Han) COLD : The CoSy Localization Database

Title : COLD : The CoSy Localization Database

Author : A. Pronobis, B. Caputo

Abstract : Two key competencies for mobile robotic systems are localizationand semantic context interpretation. Recently, vision has become themodality of choice for these problems as it provides richer and moredescriptive sensory input. At the same time, designing and testingvision-based algorithms still remains a challenge, as large amounts ofcarefully selected data are required to address the high variability ofvisual information. In this paper we present a freely available databasewhich provides a large-scale, flexible testing environment forvision-based topological localization and semantic knowledge extractionin robotic systems. The database contains 76 image sequencesacquired in three different indoor environments across Europe. Acquisitionwas performed with the same perspective and omnidirectionalcamera setup, in rooms of different functionality and under variousconditions. The database is an ideal testbed for evaluating algorithmsin real-world scenarios with respect to both dynamic and categoricalvariations.


[Full text]