基于线性矩和小波变换的医学图像特征提取

Medical Image Feature Extraction Based on L-moment and Wavelet Transform

应用小波矩作为图像特征量以及小波变换进行医学图像的自动特征提取,利用小波变换对线性矩进行分析,提取图像的各分辨率下的小波系数经快速傅里叶变换作为特征。使计算量和数据量大大减少,为特征的选取带来方便。矩特征由于其良好的稳定性、抗噪性和旋转不变性在图像识别中受到了广泛的关注和应用。由小波矩来构造目标的旋转不变性的特征,可以克服传统矩的弊端,又具有旋转不变性的特征。

Wavelet moment as an image characteristic quantity and the use of wavelet transform for medical image automatic feature extraction using wavelet transform analysis of linear moments to extract the image of the wavelet coefficients of the various resolutions by the fast Fourier transform as the feature. To calculate volume and significandy reduced the amount of data, in order to bring convenience to the selection of features. Moment feature because of its good stability, anti-noise and rotation invariance in image recognition has been widespread concern and applications. Wavelet to construct the target from the moment the rotation invariant features, can overcome the shortcomings of the traditional moments ,but also has rotation invariant features.