My voice recogniser uses some of the same techniques that are used in image processing,such as threshholding,chain coding,and perimeter estimation,and you could quite happily use a 8052 to do video image recognition as long as you were prepared to wait a while.Its just a matter of scale,it takes 1/10 of a second to process
1024 sound samples at 12Mhz,so video you need to sample at >13Mhz for colour and 312 lines,2 frames,do the maths.