An Efficient Digital Architecture for Principal Component Neural Network and its FPGA Implementation
N. Sudha, Cyril Prasanna, Veezhinathan Kamakoti · IETE Journal of Research · 2007
Principal Component Analysis (PCA) finds wide applications in machine vision. The neural network that performs PCA is called Principal Component Neural Network (PCNN). This paper presents a digital hardware design for principal component neural network. The design is efficient in the sense that the learning rule is implemented with a reusable circuit. Results of FPGA implementation of the design show that as many as 500 input vectors can be processed during training phase and 700 input vectors during retrieval phase in a second. Such results are valuable for high-speed applications.