Online Human Action Recognition Using Deep Learning for Indoor Smart Mobile Robots

Depth Sensors-Based Action Recognition Using a Modified K-Ary Entropy Classifier

“…Third, the elbow point in a human silhouette has been calculated by taking the center of shoulder and hand points [58]. The elbow point on a human's arm has been depicted in equation (6).…”

“…Moreover, human actions recorded with a various sensors, including depth sensors, smartphone sensors, RGB sensors, and others, to perform HAR are usually sensitive to changes in lighting and background clutter. Furthermore, it is impractical to use many cameras to achieve HAR [6]. Thus, with the recent advancement in vision-based technology, depth-based sensors, such as low-cost Kinect, have improved a lot in efficiency and quality.…”