Human action analyses and recognition are challenging problems due to large variations in human motion and appearance, camera viewpoint and environment settings. Among those methods, some methods get significant attention from many researchers in the computer vision field due to their better robustness and performance.