Identification research of pulmonary nodules based on PCA and the improved BP neural network

摘要

We propose a detective method of pulmonary nodules which is applied to CT image based on PCA and the improved BP neural network. Characteristics of nodules and vascular which are processed by PCA have been obtained, classified by applying the improved BP neural network, selected by competition, recognized by results. Experiment results show that implementation of the recognition model is simple, recognition rate is high, speed of training and recognition is fast, and the trained network has very good classification ability.