- 作者: Ramasamy Ramachandran, Natarajan Gunasekaran
- 中文摘要: The Hopfield model with bipolar neural states (-1,1) has an advantage in that the output signal-to-noise ratio is times that of the unipolar neural states (0,1). Also, the information storage capacity of the bipolar neural states can be twice that of the unipolar model. It is difficult to represent bipolar quantities in an intensity distribution. A method to achieve full bipolar performance in a single channel optical associative memory is presented in this paper. A two dimensional bipolar Hopfield model optical neural network has been implemented by coding the biased interconnection weights, a distributed background and an input-dependent dynamic threshold on a single mask. Content addressability properties are improved through the introduction of a distributed background. Computer simulations of two dimensional bipolar neural networks have been performed.
- 英文摘要: --
- 中文關鍵字: associative memory, neural networks, image processing, pattern recognition, optical computing.
- 英文關鍵字: --