Real-time estimation of human body postures from monocular thermal images

This paper proposes a new real-time method for estimating human body postures from thermal images acquired by an infrared camera, regardless of the background and lighting conditions. Distance transformation is performed for the human body area extracted from the thresholded thermal image, in order to calculate the center of gravity. After the orientation of the upper half of the body is obtained by calculating the moment of inertia, significant points such as the top of the head and the ends of the hands and feet are heuristically located. In addition, the elbow and knee positions are estimated from the detected (significant) points, using a genetic-algorithm-based learning procedure. The experimental results demonstrate the robustness of the proposed algorithm and real-time performance (faster than 20 frames per second).