NET32K high speed image understanding system

Two NET32K neural-network chips are integrated on a board system with an SBus interface, to serve as a high speed image analysis platform. The system is optimized for convolutional networks. Up to 64 Kernels of size 16×16 pixels are scanned simultaneously over an image. In this way, simple geometric shapes are extracted from an image, representing its content in a compact form. A standard processor can then do the high level interpretation. To prevent I/O bottlenecks between board and host, several high speed programmable logic devices are included on the board to implement tapped delay lines and compression/decompression algorithms. The board can process 20 frames per second, achieving over 100 GC/s (billion connections per second). The SBus interface makes it possible to directly “plug” the board into a SUN Sparcstation, providing a compact and low cost solution for complex image analysis tasks. Several document processing applications are described