一种改进的3维块匹配视频去噪算法

An Improved Video Denoising Algorithm Based on 3D Block Matching

针对Video Block-Matching and 3-D Filtering(VBM3D)算法耗时高且去噪视频存在块效应的问题进行了改进。在对变换域系数进行收缩时,采用连续阶导数阈值法,代替原算法中的硬阈值法,减少块效应;在帧内匹配时,采用基于积分图思想的图像块距离计算加速方法;在帧间匹配时,使用帧间预测性匹配方法,减少计算量,提高算法效率。理论分析和实验结果表明,改进后的算法不仅能有效改善原VBM3D算法中的块效应,且算法复杂度大大降低。

The video block-matching and 3-D filtering(VBM3D) algorithm was improved,which had block effect and low-efficiency.A continuous derivative threshold method was introduced during the shrinkage of the coefficient in transform domain,replacing the hard threshold method.A fast method based on the integral images was used to calculate the distance between image blocks in intra frames matching.An improved searching method of predictive-search block-matching for inter frames matching was presented.The computation was reduced and the efficiency of the algorithm was improved on the premise of finding similar blocks.The experimental results and the theoretical analysis demonstrated that the improved method achieved excellent denoising performance.The result of block effect in the original VBM3 D algorithm was improved and the computation complexity of improved algorithm was reduced greatly.