Artifical retina devices-fast front ends for neural image processing systems

As more and more fast hardware neural networks become available, greater attention should be given to the sensory front ends of these networks. In this paper, we introduce an artificial retina device and show how it can be used as a visual front end for neural networks. Our device can perform many different preprocessing operations on projected images, it features a high processing speed, and it has a parallel output port which can easily be connected to a hardware neural network. Experimental results demonstrating the performance of systems for feature extraction, image compression, and the recognition of characters from a phonetic Japanese alphabet are also provided.

[1]  D. Hammerstrom,et al.  A VLSI architecture for high-performance, low-cost, on-chip learning , 1990, 1990 IJCNN International Joint Conference on Neural Networks.

[2]  S. Tam,et al.  An electrically trainable artificial neural network (ETANN) with 10240 'floating gate' synapses , 1990, International 1989 Joint Conference on Neural Networks.

[3]  M. Takahashi,et al.  Optical neurochip with learning capability , 1992, IEEE Photonics Technology Letters.

[4]  Ulrich Rückert,et al.  VLSI Design of Neural Networks , 1990 .

[5]  John Wawrzynek,et al.  SPERT: a VLIW/SIMD neuro-microprocessor , 1992, [Proceedings 1992] IJCNN International Joint Conference on Neural Networks.

[6]  John Wawrzynek,et al.  SPERT: a neuro-microprocessor , 1995 .

[7]  K. Kyuma.  Optical neuro-devices , 1993 .

[8]  J. Beichter,et al.  Design of a 1st Generation Neurocomputer , 1991 .

[9]  K Kyuma,et al.  Variable-sensitivity photodetector that uses a metal-semiconductor-metal structure for optical neural networks. , 1991, Optics letters.