2 June 2020 in IEEE Access
Falls have caught great harm to the elderly living alone at home. This paper presents a novel visual-based fall detection approach by Dual-Channel Feature Integration. The proposed approach divides the fall event into two parts: falling-state and fallen-state, which describes the fall events from dynamic and static perspectives. Firstly, the object detection model (Yolo) and the human posture detection model (OpenPose) are used for preprocessing to obtain key points and the position information of a human body. Then, a dual-channel sliding window model is designed to extract the dynamic features of the human body (centroid speed, upper limb velocity) and static features (human external ellipse). After that, MLP (Multi-layer Perceptron) and Random Forest are applied to classify the dynamic and static feature data separately. Finally, the classification results are combined for fall detection. Experimental results show that the pro-posed approach achieves an accuracy of 97.33% and 96.91% when tested with UR Fall Detection Dataset and Le2i Fall Detection Dataset.
1 December 2015 in Journal of Computational and Theoretical Nanoscience
Face tracking is one of the most fundamental problems to be addressed by computer vision due to that it is the precondition of almost all the face-based applications. To address this problem, inspired by the mechanism of human vision system in face tracking, this paper proposes a human body fuzzy tracking based method in which it not only takes advantage of both the moving trajectory and the SIFT features matching to track the target faces indirectly, but also introduces a tracked-thenlearning feature updating mechanism to improve its robustness greatly. Experimental results prove that our method has sound efficiency and excellent robustness on most of the unfavorable factors, such as cosmetic, camouflage, the changes of head pose or facial expressions.