基于颜色和纹理特征的胶囊内镜图像分类

Capsule endoscope image classification based on color and texture features

针对常用的计算机辅助分析方法受消化道气泡、光照和拍摄角度等影响难以对胶囊内镜图像取得较好的分类效果的问题,提出一种结合颜色矩、小波变换和共生矩阵的特征提取方法,并用SVM将图像分为健康和病变两类。计算HSV空间去噪后图像的颜色矩,同时采用小波变换选择中高频带后重构图像并计算其共生矩阵特征值,将提取出的特征值归一化,作为SVM的输入进行训练和分类。实验结果表明,该方法正确率可达98.88%,相比其他方法取得了更好的分类结果。

The commonly-used computer aided analysis method is difficult to obtain a better classification result of capsule endoscopy images due to the influence of digestive tract bubbles,illumination and shooting angle.Therefore,a feature extraction method based on color moment,wavelet transform and co-occurrence matrix is proposed,and the SVM is used to classify the images into healthy and diseased categories.The color moments of the denoised images in the HSV space are calculated,and the wavelet transform is used to select the middle and high frequency bands to reconstruct the images and calculate the feature values of their co-occurrence matrixes.The extracted feature values are normalized as the inputs of SVM for training and classification.The experimental results show that the correct rate of the proposed method can reach up to 98.88%,which is much better than other methods.