Studies On Image Recognition Based On Wavelet Transform And Artificial Neural Network

The purpose of this thesis is to utilize the predominance of the wavelet transform and artificial neural network for the recognition of images.This paper revolves around the central task of image identification. It is mainly about collecting and preprocessing the original data of target images, methods of invariable feature extraction and the identification technology of the artificial neural network.At the preprocessing part, we smooth and denoise the target image firstly, then we detect its edge and buildup it, lastly, in order to eliminate the effect of the translation, scaling, skewing and rotation on the recognition result. We proposed a method to normalize the target image, which involved in the feature extraction part.During the feature extraction part, we studied two methods: one is the Zernike moment invariant method, the other method is based on the wavelet transform. By using both of the two methods, we extract the features of the target images, then we applied the features to the artificial neural networks in the recognition part.We utilize the artificial neural networks method to identify the images in the recognition part. An improved back propagation (BP) algorithm was introduced to train the neural network for recognition. We trained and identified the images by the improved BP algorithm with the data obtained from both Zernike moment invariant method and the wavelet transform method. Simulation results show that the method proposed in this thesis is well validate for recognizing objects and the eigenvalue extracted by thewavelet method have a more effect on the identification of images than by the method of Zernike moment invariant.