一种基于自适应颜色压缩的织物图案提取方法

Adaptive Color-reduction Based Draft Extraction of Textile Image

文章介绍了一种新的基于主成分分析网络PCA与自组织特征映射网络SOFM的颜色自适应压缩技术,并首先将该技术应用到了纹理状织物图案的提取研究中。研究表明:提取主颜色系对消除图像中的噪声污染成分非常有效,能够减少原图像上因噪声干扰引起的纹理分割误差。通过竞争学习的分类识别引入了像素间的空间位置关系,使得经神经网络分类出的图像与原图像最接近,并提高了在图案提取分割中的计算效率。该研究结果在获取织物图案设计与仿制中可减少人机交互量,提高识别效率,在工程应用上具有相当的实用价值。

Reduction of the number of image colors is crucial for many applications,such as segmentation and presen-tation of color images.The adaptive color reduction technique is used in this paper to deal with the draft extraction of textile image,which is based on a Neural Network(NN)structure using a tree clustering procedure.In each node of the tree,a self-organized Neural Network Classifier(NNC)is used which consists of a Principal Component Analyzer(PCA)and a Kohonen Self-organized Feature Map(SOFM)neural network.The training set of the NN consists of the image color values and additional spatial features extracted in the neighborhood of each pixel and the output neurons of the NNC define the color classes for each node.Therefore,the final image not only has the proper colors,but its structure approxi-mates the local characteristics used.Using the adaptive procedure and different local features for each level of the tree,the initial color classes can be split or merged even more.This research shows that it can be considered as a technique for the detection of the number of dominant colors in an image and is suitable for eliminating uncommon colors in an image by examining only the contribution of the image colors in the color space.Several experimental and comparative results,exhibits the performance of the technique to the mixed type draft extracting application of textile or fabric im-age.