基于树型小波变换和Hopfield神经网络的纹理图像分割

Texture Image Segmentation Based on Tree Structure Wavelet Transformation and Hopfieid Neural Network

纹理分割是图像分析、模式识别、计算机视觉等领域的基础,也是个经典难题.采用树型小波变换方法提取纹理图像特征,并采用Hopfield神经网络进行象素聚类,从而实现对纹理图像的分割,实验结果表明该方法是有效的.

Texture image segmentation is foundation of image analysis, model recognition and computer vision, which is also a classic problem. The texture features are extracted by using tree structure wavelet transformation, and then Hopfield neural network is used to cluster pixels to realize segmentation of the texture images. The experimental results show that the method is effective.