基于图像块先验的低秩近似和维纳滤波的去噪算法

利用混合高斯模型(gaussian mixture model, GMM)学习自然图像块的纹理结构,提出一种基于图像块先验的低秩近似和维纳滤波的去噪算法。该算法能够同时利用外部图像块的先验结构信息和内部图像的自相似性,对待去噪图像进行分块聚类,并根据每类相似块的数量进行协同滤波。当相似图像块数量较多时,采用低秩近似的方法复原,有效利用图像的内部自相似性;当相似图像块数量较少时,采用维纳滤波,利用先验信息保持图像重要的纹理结构。试验结果表明此方法较适用于弧形边界和角点等存在较少相似块的自然图像,其峰值信噪比(peak signal to noise ratio, PSNR)和视觉效果优于目前部分主流算法。

Image patch prior based denoising algorithm by using low rank approximation and Wiener filtering

A Gaussian mixture model(GMM)was used to study the texture structure of natural image patches, and a low-rank approximation and Wiener filtering algorithm based on image patch prior were proposed. The proposed method divided the image into a number of overlapped patches and clustered them for collaborative filtering by using the prior structures of external image patch and internal image self-similarity. By grouping nonlocal similar patches, low-rank approximation was used as collaborative filtering to recover the texture structures. When the number of similar patches was small, Wiener filtering with patch prior was adopted to preserve texture features. The experimental results indicated that the proposed method was more suitable for the images with fewer similar patches like boundary and corner etc., and showed very competitive performance with state-of-the-art denoising method in terms of Peak Signal to Noise Ratio(PSNR)and visual quality.