신경회로망의 하드웨어 구현 및 암의 패턴 분류 = Hardware Implementation of Neural Network and its Application

서명/저자: 신경회로망의 하드웨어 구현 및 암의 패턴 분류 = Hardware Implementation of Neural Network and its Application / 동성수

초록/해제: 요약 :Hardware implementation of neural network has difficulties in the extension of synapses and the programmability for relocating neurons. In this paper, the architecture of a new hardware is proposed to solve these problems. The proposed hardware based on traditional SIMD enables connections of network to be dynamically and easily reconfigured without synthesizing and mapping original design for each use. Using additional modular processing unit the number of neurons and synapses increases. To show the extensibility of the new structure, Kohonen network is formed and tested. The performance comparison with C-model simulation shows its superiority in the aspects of performance and flexibility.