Classification of human body parts using histogram of oriented gradients
R. Newlin Shebiah, A. Aruna Sangari · 2019
In computer vision, classification of human body parts faces challenges such as partial occlusion of body parts, viewpoint variation, lightning and appearance changes, etc. In this paper in an outdoor scene human is detected using Histogram of Oriented Gradients (HOG) and Cascade adaboost classifier. Further the detected human is parsed into regions like head, torso, leg and further subdivided into neck, left-right elbow, left-right shoulder, left-right wrist, left-right hip, left-right ankle and left-right knee. Each distinct part was trained and classified using Histogram of Oriented Gradients (HOG) and Support Vector Machine (SVM) classifier. The proposed method was evaluated using Leeds Sports Pose (LSP) dataset.