Some pioneers of computer science, way back in the golden age of cybernetics, believed that the age-old mysteries of the mind would be resolved in the near future. Their optimism was founded on the rapid advances in computer hardware and the new discoveries of biology, biochemistry and neuroanatomy. As we know now, these advances have continued at an ever-accelerating pace. Computers are now faster by six orders of magnitude, and digital storage capacities increased by at least six orders, too. On the other hand, experimental neuroscience has advanced far beyond what was known when I went to school. The biochemical cascade from photon detection, ion channels, action potentials, up to the various brain areas that are involved in vision is known up to astonishing level of detail.
Yet there is still a gap.
We are still not able to build artificial brains. No machine can run as well as a dog, is as curious as a rat, understands speech as well as a child. Hey, heck, we can’t even sort a heap of photographs of cats and dogs ( Project Asirra ). Our robots are slow zombies, lots of sensors but still deaf and blind. The robots of North Central Positronics™ are nowhere to be seen.
This is odd. Extremely odd. Uncomfortably so.
Of course, there has been progress. There are now commercial systems for optical character recognition, continuous speech recognition, and even military robots are being sold. But these systems are not new. They were first demonstrated decades ago. Advanced hardware gradually made old approaches more viable, but many of the early shortcomings still remain with us after decades of research.
Commercial OCR systems are for the most part still ignoring Neural Networks or Support Vector Machines ( „based on old technology“ ). The Ocropus Project attempts to build a state of the art system.
Update: Google offers image search with face recognition (here). Obviously, I’ve been ignoring some recent progress. Google acquired Neven Vision in 2006.
EP1072018 : Wavelet-Based Facial Motion Capture for Avatar Animation 1072014 Face Recognition from Video Images 218457 Face Recognition from Video Images 218458 Wavelet-Based Facial Motion Capture for Avatar Animation EP1072018 Wavelet-Based Facial Motion Capture for Avatar Animation 1072014 Face Recognition from Video Images 6714661 Method & System for Customizing Facial Feature Tracking Using Precise Landmark 6222939 Labeled Bunch Graphs for Image Analysis (EYEM1160/ NE01) 6356659 Labeled Bunch Graphs for Image Analysis 6563950 Labeled Bunch Graphs for Image Analysis 6466695 Procedure for Automatic Analysis of Images & Image Sequences Based on Two Dimensional Shape Primitives 6272231 Wavelet-Based Facial Motion Capture for Avatar Animation 6580811 Wavelet-Based Facial Motion Capture for Avatar Animation 6301370 Face Recognition from Video Images