一种基于关联向量回归的盲图像复原

Blind Image Restoration Based on Relevance Vector Regression

提出一种基于关联向量回归模型的盲图像复原算法。用模糊噪声图像作为训练集合,由此得到关联向量回归模型。该模型可用于恢复受不同类型模糊和不同噪声污染的图像。实验结果表明:同其他盲复原算法相比,该算法对点扩展函数(PSF)类型和噪声都有较强的鲁棒性。同时,从改善信噪比大小和主观视觉两个方面,该算法都能取得良好的复原效果。

This paper proposed a new blind image restoration algorithm based on relevance vector regression （RVR） models. Using blurred and noise image as training sets,RVR models are developed. The models can then be used to recovery different images corrupted by different types blur and noise at different levels. In contrast to other blind restoration algorithms ,experimental results show that this algorithm is very robust in regard to the blurs type and noise level. At the same time,the algorithm can obtain excellent result in the peak signal-to-noise rations （PSNR） and subjective visual effect.