基于块匹配稀疏下医学视频数据降噪算法研究

Research on medical sparse video denoising algorithms based on block matching

针对现代医疗临床诊断和医学研究对医学视频降噪的需求,通过医学视频数据的帧拆解,充分研究了医学视频在变换域中的稀疏特性,基于三维块匹配算法(Blocking Matching 3D,BM3D)实现了对医学视频数据的去噪声处理。由于在应用时充分考虑到了医学图像在变换域中的稀疏性,本算法克服了BM3D算法高时间复杂的缺陷。通过与其他降噪算法的对比分析,本算法去噪后的图像和原图像间具有最高的相关性。

In view of the need of denoising of medical video,In this paper,the sparsity of medical video in transform domain is fully studied by frame decomposition of medical video data.The denoising of medical video data is realized by using the(Blocking Matching 3 D,BM3 D) algorithm.Because the sparsity of medical images in transform domain is fully considered in application,this algorithm overcomes the shortcoming of high time complexity of BM3 D algorithm.Compared with other denoising algorithms,the denoised image has the highest correlation with the original image.