Meet the cousin of face recognition. Recognition of clothing categories from videos is appealing to emerging applications such as security, intelligent customer profile analysis and computer-aided fashion design. We have developed a complete system to tag clothing categories in real-time, which addresses some practical complications in surveillance videos.
Specifically, we take advantage of face detection and tracking to locate human figures and develop an efficient clothing segmentation method utilizing Voronoi images to select seeds for region growing. We compare clothing representations combining color histograms and 3 different texture descriptors.
Clothing recognition is an advanced image processing application, which may benefit customer profile analysis, context-aided people identification and computer aided fashion design. Although this problem attracts increasing research interests in recent years, a real-time clothing recognition system, especially for surveillance videos, remains challenging, primarily due to two reasons.
First, such a system involves a series of difficult sub-problems including face detection and tracking, human figure or clothing segmentation, and effective clothing representations. Second, the differences among various clothing categories are inherently subtle and even vague for human, thus considerable computations are required to discern them. However we continue to make advances in these areas, and the overall clothing recognition model is constantly improving.