基于贝叶斯最小均方差的OCT图像散斑处理

Speckle processing for OCT image based on Bayesian least mean square error criterion

对于光学相干层析图像中的散斑噪声,提出了一种基于贝叶斯的降噪算法.通过将带有噪声的成像数据放在对数空间中,从噪声的高斯分布中抽取样本,根据图像相邻像素之间的相关性,对样本内的像素赋予相应的权值,用加权直方图估计后验分布,并运用一般贝叶斯最小均方差得出图像的无噪声数据.该算法与传统的小波变换降噪和中值滤波去噪相比,在信噪比(SNR)和等效视数(ENL)方面都有明显的改善,在一定程度上提升了图像质量.

This paper presents a noise reduction algorithm for the speckle noise in optical coherence tomography images based on Bayesian criterion. First ,the noisy imaging data is put into the logarithmic space and sample is extracted from the data with noise of Gaussian distribution. Then pixels within the sample are given relevant weights based on the correlation between adjacent pixels in the image. Finally, the posterior distribution is estimated by using a weighted histogram approach and the noise-free data is esti- mated using generic Bayesian least mean square error estimate. Compared with traditional wavelet transformation noise reduction and median filtering denoising, this method obviously improves the signal-to-noise ratio （SNR） and the equivalent apparent num- ber （ENL） of OCT image. Thus the image quality is enhanced to some extent.