Physorg reports on work by the McGovern Institute for Brain Research at MIT that supports a previously held suspicion that the human brain has a separate area that specializes only in recognizing faces. They also found that another area, immediately adjacent to the facial recognition area, performs the same task for bodies (and not faces). The researchers targetted this particular region of the brain for study because of observations that people with physical damage to it lost the ability to recognize faces. The researchers used a next-generation imager with increased resolution to distinguish between the two distinct areas involved. The work was published in The Journal of Neuroscience (23 November 205).
The results of this research seem to be in agreement with the parallel tracks that machine-recognition of people in surveillance videos is taking - namely, recognition of the face and recognition of the gait. Currently, face recognition works pretty well if the entire face is clearly visible (i.e. sufficient resolution, unobscured, full face towards the camera, and well lit), but falls apart if it isn't. In uncontrolled situations, you can guess what usually happens (Murphy's Law applies here). This has led to efforts to develop new image processing algorithms that recognize a body or its actions, such as gait.