Detection and Tracking of Humans by Probabilistic Body Part Assembly

Abstract

This paper presents a probabilistic framework of assembling detected human body parts into a full 2D human configuration. The face, torso, legs and hands are detected in cluttered scenes using boosted body part detectors trained by AdaBoost. Body configurations are assembled from the detected parts using RANSAC, and a coarse heuristic is applied to eliminate obvious outliers. An a priori mixture model of upper-body configurations is used to provide a pose likelihood for each configuration. A joint-likelihood model is then determined by combining the pose, part detector and corresponding skin model likelihoods. The assembly with the highest likelihood is selected by RANSAC, and the elbow positions are inferred. This paper also illustrates the combination of skin colour likelihood and detection likelihood to further reduce false hand and face detections.

Detecting and segmenting humans in crowded scenes

“…Subsequently, inferences are made with respect to the best assembly of existing part hypotheses. Approaches such as AdaBoost have been used with some degree of success to learn body part detectors such as the face [10], hands, arms, legs, and torso [5] [8].While this class of approaches is attractive, detection of parts is itself a challenging task. This is particularly difficult in the class of scenes in which we are interested, which consist of crowded scenes containing significant occlusion amongst many parts.…”

“…Another family of approaches models humans as a collection of parts [5], [7], [8] and [9]. Typically this class of approaches relies on a set of low-level features which produce a series of part location hypotheses.…”

People Detection Using Color and Depth Images

“…Prompted in part by security considerations [20], new techniques, protocols and standards have emerged in the past decade. Some approaches have used silhouettes [5] or body-part matching [14,21,23,22]. The combination of cascades of increasingly complex classifiers has produced fast and robust recognition algorithms [28] for relatively stylized person poses.…”

Probabilistic 3D Tracking: Rollator Users' Leg Pose from Coronal Images

“…[23]), colour (e.g. [13]) and anthropometric symmetry have been exploited in several works. However, these low-level cues along with edges, corner features (ex.…”