Human detection using a combination of face, head and shoulder detectors
Feng Su, Gu Fang, Ju Jia Zou · 2016
Human detection is an active topic with many interesting applications in computer vision. This paper proposes an improved human detection method to extract and segment the human body figure from a given image without prior information. This is done by detecting face, head and shoulder separately, face detection is obtained from Haar classifier while head and shoulder detections are determined from gradient maps, the final human body segmentation is generated based on gap detection and golden ratio. Experiment Results have shown that this method is capable of detecting and segmenting the human body figures robustly from various images.