A real-time embedded system for human action recognition using template matching

Human action recognition system proposed here recognizes the behavior of a person in real-time. The system aims at communicating the recognized gestures with the camera system. The proposed system initiates on identifying the human action provided in the database and it sends the trigger signals to the camera system so as to record and store the video stream within the system. The video stream is then transmitted to through web page with the aid of IoT. The system is also provided with a buzzer that activates on recognizing the human action. Template matching algorithm is used to recognize and identify actions from the captured video frame directly. The system is realized using Beaglebone board with ARM CORTEX A8 processor and OpenCV is used as the software tool to implement the system.