Research of X Ray Image Recognition Based on Neural Network

Abstract

In order to make x ray image recognition to apply to BP neural network effectively. Firstly, preprocess x ray image and k-mean segmentation image, extract image feature, The extracted feature as input of BP neural network for training and testing network. Standard BP neural network and improved BP neural network are used to recognize x ray image of this paper, then compare learning rate, training error and recognition rate of two algorithms. Innovation is using improved BP network model to detect the target object. It can successfully detect the object of x ray image with covered by other object or without occlusion. Experiment shows that improved BP neural network has faster learning rate, less error, high recognition rate, it can identify and detect the target object of x ray image effectively.