基于DRBM和边缘检测的脑部磁共振图像分类

Classification of brain MRI images based on DRBM and edge detection

为提高图像分类精度,文中提出了一种DRBM网络结合边缘检测的图像分类方法。该方法首先进行边缘检测,然后将提取的纹理叠加到原图中,实现图像纹理加强。接着构建可视层-隐层-分类器三层DRBM,实现特征提取并分类。实验证明,相比较传统基于单一特征的分类方法,文中方法取得了较高的分类准确率,具有更好的图像分类性能。

In order to improve the accuracy of image classification,a new method edge detection combined DRBM network is proposed. Firstly,the edge of the image is detected,and the texture feature extracted is superimposed into the original,enhancing the image texture. Then a three-layer DRBM consist of a visual layer,a hidden layer and a classifier is built,extracting feature and classifying images. The experiment shows that the method gets a higher classification accuracy and better image classification performance compared with the traditional classification method which based on single statistic feature.