Digital Recognition Based on Neural Network and FPGA Implementation

In recent years, neural network is widely used in many fields such as pattern recognition, signal processing and so on. Digital recognition is an important branch in the field of pattern recognition. The main content of this paper is the realization of digital recognition of BP neural network on the hardware. Because in the actual scene, many times in the system design, we need to consider the reality. Compared with the traditional software implementation, FPGA has the characteristics of parallelism, which can greatly improve the processing speed and meet the application requirements. In this paper, the off-line training method is used to store the off-line training data into RAM, and the forward propagation process of BP neural network is realized on FPGA.