Texture Image Segmentation Method Based on Wavelet Transform and Neural Networks OK外语英文电子书.pdfVIP

Texture Image Segmentation Method Based on Wavelet Transform and Neural Networks Jing Zhang Shunichiro Oe Information Science Intelligent Systems University of Tokushima, Japan 2-1. Minami-Josaniima. Tokushima. 770-8506 JaDan ahang@is.tokushimi-u.ztc.jp oe@is.tokushima-u.ac.jp ABSTRACT texture features can be divided into statistical and structural methods. Structural methods describe This paper presents an effective texture image seg- textures in terms of texture primitives and spatial mentation algorithm by using wavelet decomposition relationships among them. Now many computa- and band-pass neural networks. This approach is ap- tional methods have been proposed and applied to plied to segment an random texture image into sev- extract texture feature. Commonly used texture fea- eral homogeneous areas. The basic idea of proposed tures include Fourier domain energy, co-occurrence method is first decomposing an original image into matrices, texture energy measures, Markov random several filtered images which contain information in field models, second-order gray level statistics, two- different orientation and frequency ranges, and these dimensional autoregressive model and fractal dimen- filtered images are of the same size as the origi- sion etc.[l]. But in most early work, the texture nal images. Then the ze