Building Roadmaps of Minima and Transitions in Visual Models

Sminchisescu, Cristian; Triggs, Bill

doi:10.1023/b:visi.0000042935.43630.46

Cited by 29 publications

(22 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The task of tracking a human body in monocular images is commonly carried out by fitting articulated kinematic models representing the appearance of a person to mono-or multiocular images (Sidenbladh et al, 2000;Sigal et al, 2004;Sminchisescu and Triggs, 2005). The surveys of Gavrila (Gavrila, 1999), and Moeslund and Granum (Moeslund and Granum, 2001) provide a good overview of the topic of model based 3D body tracking.…”

Section: Introductionmentioning

confidence: 99%

AUTOMATIC INITIALIZATION FOR BODY TRACKING - Using Appearance to Learn a Model for Tracking Human Upper Body Motions

Schmidt

Castrillón-Santana

2008

Proceedings of the Third International Conference on Computer Vision Theory and Applications

View full text Add to dashboard Cite

Abstract:Social robots require the ability to communicate and recognize the intention of a human interaction partner. Humans commonly make use of gestures for interactive purposes. For a social robot, recognition of gestures is therefore a necessary skill. As a common intermediate step, the pose of an individual is tracked over time making use of a body model. The acquisition of a suitable body model, i.e. self-starting the tracker, however, is a complex and challenging task. This paper presents an approach to facilitate the acquisition of the body model during interaction. Taking advantage of a robust face detection algorithm provides the opportunity for automatic and markerless acquisition of a 3D body model using a monocular color camera. For the given human robot interaction scenario, a prototype has been developed for a single user configuration. It provides automatic initialization and failure recovery of a 3D body tracker based on head and hand detection information, delivering promising results.

show abstract

Section: Introductionmentioning

confidence: 99%

AUTOMATIC INITIALIZATION FOR BODY TRACKING - Using Appearance to Learn a Model for Tracking Human Upper Body Motions

Schmidt

Castrillón-Santana

2008

Proceedings of the Third International Conference on Computer Vision Theory and Applications

View full text Add to dashboard Cite

show abstract

“…Only few authors have addressed the problem of 3D full-body tracking using a single uncalibrated camera [16,15,18]. One such approach for tracking a detailed 3D human body model was proposed by Sidenbladh [16,15].…”

Section: Related Workmentioning

confidence: 99%

“…To cope with the huge search space, motion priors are used to predict the 3D body configuration prohibiting the tracking of unconstrained motions. Following Sidenbladh's work, Sminchisescu used a more precise modeling of the 3D body model and a complex parameter space exploration [18], but the computational time required prohibits its use for realtime tracking. To cope with a large parameter space, kernelbased Bayesian filtering has been proposed to track objects or isolated body parts in the 2D image space (e.g., [7,3]).…”

Section: Related Workmentioning

confidence: 99%

“…Such approaches require image disparities to successfully carry out the computationally expensive depth calculation (e.g., [12]). A different approach is taken in model-based tracking methods performing a probabilistic integration of image cues (see, e.g., [16,18,4,17,13,9]). However, extracting 3D body configurations from 2D image data is connected with complicated modeling as well as feature extraction difficulties.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Kernel Particle Filter for Real-Time 3D Body Tracking in Monocular Color Images

Schmidt

Fritsch

Kwolek

7th International Conference on Automatic Face and Gesture Recognition (FGR06)

View full text Add to dashboard Cite

show abstract

“…The No Free Lunch Theorems for search and optimization discussed in Section 2.2 explains that, roughly speaking, if a search algorithm outperforms another search algorithm for a particular problem, then it will have worse performance over all other problems. Following this, this thesis follows the methodology used most predominantly throughout the works of Sminechisecu and Triggs [86,87,[89][90][91][92][93], whereby the study of the properties / topology of the probability distributions to be searched is of critical importance, and the design of search / optimization techniques based around these properties. A good example of this philosophy is Isard and Blakes CONDENSA-TION [48] algorithm, which proved very successful and brought much attention to this field.…”

Section: Aimmentioning

confidence: 99%

Techniques for binocular markerless visual tracking of 3D articulated bodies

Smith¹

View full text Add to dashboard Cite

This thesis advances methods for performing markerless visual tracking of articulated bodies using one or two cameras. The research presented aims to improve upon existing Bayesian inspired tracking methods, by examining the 'building blocks' of these tracking algorithms, in particular the measurement function design, the state space selection, and local optimization methods. Results presented in this thesis show that improvements can be made in all of these areas. These improvements are applicable to a variety of Bayesian tracking algorithms.This thesis begins by examining literature relevant to the visual tracking problem. This includes the measurement functions used by other authors, focussing on the edge detection methods used in both tracking and segmentation problems. A general overview of the global search problem is given next, as a global search is a fundamental part of a Bayesian tracking algorithm. The combination of Newton like local optimization methods and the measurement functions used in visual tracking problems is examined next, and it is shown that Newton optimizers are not ideally suited to these measurement functions. The Bayesian tracking framework is then detailed, along with a review of several existing Bayesian tracking algorithms. Finally some non Bayesian tracking algorithms are discussed.Following the literature review, details of the models used in the experiments presented in this thesis are given. These include the articulated human body model, the camera model, image gradient metrics, self occlusion treatment, and a generic colour based region measurement method.The use of graph based approaches for edge measurements is then investigated. Graph based methods are commonly used in image segmentation problems, however have not been applied to visual tracking problems. A novel method for performing edge measurements using the 'shortest path' around the object's occluding contour is presented. Unlike in the segmentation problem, self occlusion models mean the weights or costs of some graph vertices can not be determined. Different treatments for occluded graph vertices are given and evaluated. It is shown that the graph based approach produces observational likelihoods that are more accurate and have significantly fewer local maxima than the edge measurement schemes previously used in tracking problems. While this approach is computationally more expensive than other methods, it is argued that this is offset by the reduced computational expense of the global search procedure used in tracking algorithms.The choice of state space used in the tracking problems is examined next. While most authors have used a state space based on the joint angles of the human body, a Cartesian state space based on the world coordinates of limbs is proposed. While Cartesian based state spaces have been used by iii other authors for representations of kinematic models, to the author's knowledge they have not been used for full kinematic models. It is shown that that the more linear relationship betwee...

show abstract

Building Roadmaps of Minima and Transitions in Visual Models

Abstract: Becoming trapped in suboptimal local minima is a perennial problem when optimizing visual models

Cited by 29 publications

References 53 publications

AUTOMATIC INITIALIZATION FOR BODY TRACKING - Using Appearance to Learn a Model for Tracking Human Upper Body Motions

AUTOMATIC INITIALIZATION FOR BODY TRACKING - Using Appearance to Learn a Model for Tracking Human Upper Body Motions

Kernel Particle Filter for Real-Time 3D Body Tracking in Monocular Color Images

Techniques for binocular markerless visual tracking of 3D articulated bodies

Contact Info

Product

Resources

About