Action recognition in still images using a combination of human pose and context information | Synapse