Pictorial Structural Models for Human Detection in Videos

Abstract

This paper describes the detection of a human body in images. Two different approaches are used. First approach detects a human body by using a single detection window based on features of the image gradients (HOGs) and it uses a cascade of classifiers to speed up the computing time. Second approach is based on matching of pictorial structures. An articulate model of the human body is assembled from individual parts (head, torso, limbs, etc). A human body model is represented as a collection of the parts arranged in a deformable configuration. Single body parts are detected by using color characteristics, that are gained from training examples. We advance the standard implementation by detecting shapes of multiple scales. The method is accelerated by vertical symmetry of a human body. The windowed human detector is applied in order to reduce the state space. We propose a method for unsupervised learning of color appearance of the human body parts. This approach makes the detection (using matching of the pictorial structures) more robust. The method integrates the fast human detector, the pictorial structures matching and image segmentation based on graph cuts. All the used methods are tested on real datasets.

Movie

• demo.avi