Abstract

This paper investigates the recognition of human actions from three-dimensional (3-D) point clouds that encode the motions of people acting in sensor-distributed indoor environments. Data streams are time sequences of silhouettes extracted from cameras in the environment. From the 2-D silhouette contours we generate space–time streams by continuously aligning and stacking the contours along the time axis as third spatial dimension. The space–time stream of an observation sequence is segmented into parts corresponding to subactions using a pattern matching technique based on suffix trees and interval scheduling. Then, the segmented space–time shapes are processed by treating the shapes as 3-D point clouds and estimating global point feature histograms for them. The resultant models are clustered using statistical analysis and our experimental results indicate that the presented methods robustly derive different action classes. This holds despite large intra-class variance in the recorded datasets due to performances from different persons at different time intervals.

© Koninklijke Brill NV, Leiden and The Robotics Society of Japan, 2009

Overview of the approach.

Keywords: Action recognition, point cloud, global features, action segmentation

Abstract

We introduce methods for augmenting aerial visualizations of Earth (from tools such as Google Earth or Microsoft Virtual Earth) with dynamic information obtained from videos. Our goal is to make Augmented Earth Maps that visualize the live broadcast of dynamic sceneries within a city. We propose different approaches to analyze videos of pedestrians and cars, under differing conditions and then augment Aerial Earth Maps (AEMs) with live and dynamic information. We also analyze natural phenomenon (clouds) and project information from these to the AEMs to add the visual reality.

I am co-organizing the First IEEE Workshop on Computer Vision for Humanoids in conjunction with ICCV Conference in Kyoto, Japan.  This workshop will be held September 27, 2009. (9:30am – 6:00pm).

The goal of this workshop is to bring together experts from the fields of computer vision and robotics that are working on humanoid robots with vision as one of the primary modalities. Topics of interest include and are not limited to:

• Visual Learning in Robots
• Human Robot Interaction
• Grasping and Manipulation
• Learning by Demonstration
• Task Learning for Robots
• Activity Recognition and Discovery for Robot
• Humanoid Navigation in Real Environments
• Vision Devices and Systems for Robot Applications
• Application of Humanoid Robots (Indoor/Outdoor, Entertainment)

This is the first attempt at a workshop that crosses from Humanoids Research to Computer Vision Research.

The workshop includes six invited talks as well as an open poster session, where all participants are expected to present a poster describing their recent work.

Location: Kyoto University, Faculty of Engineering Bldg.#3, 2F, Room W201, in conjunction with ICCV. (See http://www.iccv2009.org/workshops/index.html).

For schedule, abstracts and other information, see the workshop website. More information about ICCV at http://www.iccv2009.org/.

Invited Speakers and Organizers after the Workshop

(Slides, NO Video)

Abstract

I will present a variety of temporal models of video that we have been studying (and developing on) for analysis and synthesis of video. Forsynthesis of videos, we have been developing representations that support example-based re-synthesis and spatio-temporal re-targeting. These approaches build on graph-based methods and we present techniques for similarity metrics for video, segmentation in video, and merging of different video streams. I will showcase a series of examples of these approaches applied to generate new videos.

For analysis of videos, we have developed a series of representations to observe and model activities in videos. Building on low-level measures of movement and motion in videos, we have incorporated higher-level temporal generative models to represent and recognize observed activities. I will discuss the strengths of a variety of State-based, Markovian, Grammar-based and Network-based representations that we have employed for recognizing activities from video. I will also discuss approaches for unsupervised discovery and recognition of activities.

Time permitting, I will describe some new efforts that move towards understanding mobile imaging and video, and video authoring and video on the web, Within these I will discuss issues of collaborative imaging, collective authoring, ad-hoc sensor networks, and peer production with images and videos. Using these concepts, to focus the conversation, I will discuss how all of these issues are impacting the field Journalism and Reporting and how we have started on a new interdisciplinary research and education effort, we call Computational Journalism.

Matthew Flagg, Atsushi Nakazawa, Qiushuang Zhang, Sing Bing Kang, Young Kee Ryu, Irfan Essa, James M. Rehg (2009), Human Video Textures In Proceedings of the ACM Symposium on Interactive 3D Graphics and Games 2009 (I3D ’09), Boston, MA, February 27-March 1 (Fri-Sun), 2009 [PDF (see Copyright) | Video in DiVx | Website ]

Abstract

This paper describes a data-driven approach for generating photorealistic animations of human motion. Each animation sequence follows a user-choreographed path and plays continuously by seamlessly transitioning between different segments of the captured data. To produce these animations, we capitalize on the complementary characteristics of motion capture data and video. We customize our capture system to record motion capture data that are synchronized with our video source. Candidate transition points in video clips are identified using a new similarity metric based on 3-D marker trajectories and their 2-D projections into video. Once the transitions have been identified, a video-based motion graph is constructed. We further exploit hybrid motion and video data to ensure that the transitions are seamless when generating animations. Motion capture marker projections serve as control points for segmentation of layers and nonrigid transformation of regions. This allows warping and blending to generate seamless in-between frames for animation. We show a series of choreographed animations of walks and martial arts scenes as validation of our approach.

Human Video Textures (Output Rendered as a Collage!)

ABSTRACT

We propose the use of 3D (2D+time) Shape Context to recognize the spatial and temporal details inherent in human actions. We represent an action in a video sequence by a 3D point cloud extracted by sampling 2D silhouettes over time. A non-uniform sampling method is introduced that gives preference to fast moving body parts using a Euclidean 3D Distance Transform. Actions are then classified by matching the extracted point clouds. Our proposed approach is based on a global matching and does not require specific training to learn the model. We test the approach thoroughly on two publicly available datasets and compare to several state-of-the-art methods. The achieved classification accuracy is on par with or superior to the best results reported to date.

Abstract

Models of activity structure for unconstrained environments are generally not available a priori. Recent representational approaches to this end are limited by their computational complexity, and ability to capture activity structure only up to some fixed temporal scale. In this work, we propose Suffix Trees as an activity representation to efficiently extract structure of activities by analyzing their constituent event-subsequences over multiple temporal scales. We empirically compare Suffix Trees with some of the previous approaches in terms of feature cardinality, discriminative prowess, noise sensitivity and activity-class discovery. Finally, exploiting properties of Suffix Trees, we present a novel perspective on anomalous subsequences of activities, and propose an algorithm to detect them in linear-time. We present comparative results over experimental data, collected from a kitchen environment to demonstrate the competence of our proposed framework.

Abstract

This paper presents an algorithm for the automatic segmentation of monocular videos into foreground and background layers. Correct segmentations are produced even in the presence of large background motion with nearly stationary foreground. There are three key contributions. The first is the introduction of a novel motion representation, “motons”, inspired by research in object recognition. Second, we propose learning the segmentation likelihood from the spatial context of motion. The learning is efficiently performed by Random Forests. The third contribution is a general taxonomy of tree-based classifiers, which facilitates theoretical and experimental comparisons of several known classification algorithms, as well as spawning new ones. Diverse visual cues such as motion, motion context, colour, contrast and spatial priors are fused together by means of a Conditional Random Field (CRF) model. Segmentation is then achieved by binary min-cut. Our algorithm requires no initialization. Experiments on many video-chat type sequences demonstrate the effectiveness of our algorithm in a variety of scenes. The segmentation results are comparable to those obtained by stereo systems.

Abstract

We present an unsupervised framework to discover characterizations of everyday human activities, and demonstrate how such representations can be used to extract points of interest in event-streams. We begin with the usage of Suffix Trees as an efficient activity-representation to analyze the global structural information of activities, using their local event statistics over the entire continuum of their temporal resolution. Exploiting this representation, we discover characterizing event-subsequences and present their usage in an ensemble-based framework for activity classification. Finally, we propose a method to automatically detect subsequences of events that are locally atypical in a structural sense. Results over extensive data-sets, collected from multiple sensor-rich environments are presented, to show the competence and scalability of the proposed framework.

Abstract

Graphical models are often used to represent and recognize activities. Purely unsupervised methods (such as HMMs) can be trained automatically but yield models whose internal structure – the nodes – are difficult to interpret semantically. Manually constructed networks typically have nodes corresponding to sub-events, but the programming and training of these networks is tedious and requires extensive domain expertise. In this paper, we propose a semi-supervised approach in which a manually structured, Propagation Network (a form of a DBN) is initialized from a small amount of fully annotated data, and then refined by an EM-based learning method in an unsupervised fashion. During node refinement (the M step) a boosting-based algorithm is employed to train the evidence detectors of individual nodes. Experiments on a variety of data types – vision and inertial measurements – in several tasks demonstrate the ability to learn from as little as one fully annotated example accompanied by a small number of positive but non-annotated training examples. The system is applied to both recognition and anomaly detection tasks.

ABSTRACT

We present an approach for tracking varying number of objects through both temporally and spatially significant occlusions. Our method builds on the idea of object permanence to reason about occlusions. To this end, tracking is performed at both the region level and the object level. At the region level, a customized genetic algorithm is used to search for optimal region tracks. This limits the scope of object trajectories. At the object level, each object is located based on adaptive appearance models, spatial distributions and inter-occlusion relationships. The proposed architecture is capable of tracking objects even in the presence of long periods of full occlusions. We demonstrate the viability of this approach by experimenting on several videos of a user interacting with a variety of objects on a desktop.

Temporal Reasoning from Video to Temporal Synthesis of Video

Abstract

In this talk, I will present some ongoing work on extracting spatio-temporal cues from video for both synthesis of novel video sequences, and recognition of complex activities. I will start off with some of our earlier work on Video Textures, where repeating information is extracted to generate extended sequences of videos. I will then describe some of our extensions to this approach that allow for controlled generation of animations of video sprites. We have developed various learning and optimization techniques that allow for video-based animations of photo-realistic characters. Then I will describe our new approach for image and video synthesis that builds on optimal patch-based copying of samples. I will show how our method allows for iterative refinement and extends to synthesis of both images and video from very limited samples. In the next part of my talk, I will describe how a similar analysis of video can be used to recognize what a person is doing in a scene. Such an analysis of video, aimed at recognition, requires more contextual information about the environment. I will show how we leverage contextual information shared between actions and objects to recognize what is happening in complex environments. I will also show that by adding some form of grammar (we use Stochastic Context Free Grammar) we can recognize very complex, multi-tasked activities.

If time permits, I will describe (very briefly) the Aware Home project at Georgia Tech, which is one primary area of ongoing and future research for me and my group. Further information on my work with videos is available from my webpage at http://www.cc.gatech.edu/~irfan

Abstract

Speech reading, also known as lip reading, is aimed at extracting visual cues of lip and facial movements to aid in recognition of speech. The main hurdle for speech reading is that visual measurements of lip and facial motion lack information-rich features like the Mel frequency cepstral coefficients (MFCC), widely used in acoustic speech recognition. These MFCC are used with hidden Markov models (HMM) in most speech recognition systems at present. Speech reading could greatly benefit from automatic selection and formation of informative features from measurements in the visual domain. These new features can then be used with HMM to capture the dynamics of lip movement and eventual recognition of lip shapes. Towards this end, we use AdaBoost methods for automatic visual feature formation. Specifically, we design an asymmetric variant of AdaBoost M2 algorithm to deal with the ill-posed multi-class sample distribution inherent in our problem. Our experiments show that the boosted HMM approach outperforms conventional AdaBoost and HMM classifiers. Our primary contributions are in the design of (a) boosted HMM and (b) asymmetric multi-class boosting.

Abstract

We present propagation networks (P-nets), a novel approach for representing and recognizing sequential activities that include parallel streams of action. We represent each activity using partially ordered intervals. Each interval is restricted by both temporal and logical constraints, including information about its duration and its temporal relationship with other intervals. P-nets associate one node with each temporal interval. Each node is triggered according to a probability density function that depends on the state of its parent nodes. Each node also has an associated observation function that characterizes supporting perceptual evidence. To facilitate real-time analysis, we introduce a particle filter framework to explore the conditional state space. We modify the original condensation algorithm to more efficiently sample a discrete state space (D-condensation). Experiments in the domain of blood glucose monitor calibration demonstrate both the representational power of P-nets and the effectiveness of the D-condensation algorithm.

Abstract

We propose a spectral partitioning approach for large-scale optimization problems, specifically structure from motion. In structure from motion, partitioning methods reduce the problem into smaller and better conditioned subproblems which can be efficiently optimized. Our partitioning method uses only the Hessian of the reprojection error and its eigenvector. We show that partitioned systems that preserve the eigenvectors corresponding to small eigenvalues result in lower residual error when optimized. We create partitions by clustering the entries of the eigenvectors of the Hessian corresponding to small eigenvalues. This is a more general technique than relying on domain knowledge and heuristics such as bottom-up structure from motion approaches. Simultaneously, it takes advantage of more information than generic matrix partitioning algorithms.

ABSTRACT

In this paper we introduce a new algorithm for image and video texture synthesis. In our approach, patch regions from a sample image or video are transformed and copied to the output and then stitched together along optimal seams to generate a new (and typically larger) output. In contrast to other techniques, the size of the patch is not chosen a-priori, but instead a graph cut technique is used to determine the optimal patch region for any given offset between the input and output texture. Unlike dynamic programming, our graph cut technique for seam optimization is applicable in any dimension. We specifically explore it in 2D and 3D to perform video texture synthesis in addition to regular image synthesis. We present approximative offset search techniques that work well in conjunction with the presented patch size optimization. We show results for synthesizing regular, random, and natural images and videos. We also demonstrate how this method can be used to interactively merge different images to generate new scenes.

ABSTRACT

We propose to develop a comprehensive framework for the joint analysis of audio-visual signals obtained from spatially distributed microphones and cameras. We desire solutions to the audio-visual sensing problem that will scale to an arbitrary number of cameras and microphones and can address challenging environments in which there are multiple speech and nonspeech sound sources and multiple moving people and objects. Recently it has become relatively inexpensive to deploy tens or even hundreds of cameras and microphones in an environment. Many applications could benefit from ability to sense in both modalities. There are two levels at which joint audio-visual analysis can take place. At the signal level, the challenge is to develop representations that capture the rich dependency structure in the joint signal and deal success-fully issues such as variable sampling rates and varying temporal delays between cues. At the spatial level the challenge is to compensate for the distortions introduced by the sensor location and pool information across sensors to recover 3-D information about the spatial environment. For many applications, it is highly desirable if the solution method is self-calibrating, and does not require an extensive manual calibration process every time a new sensor is added or an old sensor is moved or replaced. Removing the burden of manual calibration also makes it possible to exploit ad hoc sensor networks which could arise, for example, from wearable microphones and cameras. We propose to address the following four research topics: 1. Representations and learning methods for signal level fusion. 2. Volumetric techniques for fusing spatially distributed audio-visual data. 3. Self-calibration of distributed microphone-camera systems 4. Applications of audio-visual sensing. For example, this proposal includes considerable work on lip and facial analysis to improve voice communications.

Abstract

In this paper, we present techniques for recognizing com- plex, multitasked activities from video. Visual information like image features and motion appearances, combined with domain-specific information, like object context is used ini- tially to label events. Each action event is represented with a unique symbol, allowing for a sequence of interactions to be described as an ordered symbolic string. Then, a model of stochastic context-free grammar (SCFG), which is devel- oped using underlying rules of an activity, is used to provide the structure for recognizing semantically meaningful behav- ior over extended periods. Symbolic strings are parsed us- ing the Earley-Stolcke algorithm to determine the most likely semantic derivation for recognition. Parsing substrings al- lows us to recognize patterns that describe high-level, com- plex events taking place over segments of the video sequence. We introduce new parsing strategies to enable error detection and recovery in stochastic context-free grammar and meth- ods of quantifying group and individual behavior in activities with separable roles. We show through experiments, with a popular card game, the recognition of high-level narratives of multi-player games and the identification of player strate- gies and behavior using computer vision.

Recognizing Black Jack

Abstract

We present an algorithm that approximates the output of an arbitrary video processing algorithm based on a pair of input and output exemplars. Our algorithm relies on learning the mapping between the input and output exemplars to model the processing that has taken place. We approximate the processing by observing that pixel neighborhoods similar in appearance and motion to those in the exemplar input should result in neighborhoods similar to the exemplar output. Since there are not many pixel neighborhoods in the exemplars, we use techniques from texture synthesis to generalize the output of neighborhoods not observed in the exemplars. The same algorithm is used to learn such processing as motion blur color correction, and painting.