海量的数据给多媒体图像处理及计算机视觉领域带来了新的机遇与挑战。如何有效地组织利用云媒体中的数据来驱动图像的高性能处理与应用已成为学术界的一个研究热点。为了挖掘云媒体中大规模图像库对底层图像处理问题的潜在助益，本项目以云环境下图像相似度分析为切入点，构建协同利用图像内外相关性的复原模型，研究多种降质图像的复原算法。主要内容包括：针对降质图像与优质图像间质量差异带来的匹配困难，研究基于深度特征的降质-优质图像相似度分析，探索基于图像质量联合评价的优质图像筛选方法；研究基于多重约束条件的降质-优质图像块匹配方案，探索图像内外相关性的协同提升；以此为基础，针对图像CCD噪声去除、图像超分辨率、图像对比度增强等问题的特点，分别构建多重先验约束的正则化模型，并设计快速、稳健的求解算法。本项目的图像复原理论与方法可以推广到多种信号复原问题，为云媒体时代的视觉计算提供新思路与新途径。

Multimedia data on the internet is bringing new opportunities and challenges to image processing and computer vision community. How to utilize the huge database in cloud and leverage them to advance related applications in image processing is becoming a hot research topic in academia. To explore the potential assistance of the huge database in cloud to low-level image processing problems, the project proposes degraded image restoration by exploring the similarity among cloud images and building restoration models to integrate the advantages of external and internal correlations. The main research plan includes: 1) Considering the image matching difficulty caused by the quality difference between degraded images and high-quality images, investigate deep learning based image similarity analysis between degraded images and high-quality images, and investigate quality co-assessment of retrieved images; 2) Investigate the matching method between degraded patches and high quality patches using multiple constraints, and explore cooperative improvement of external and internal correlations; 3) Construct the multi-prior regularized restoration models for CCD noise removal, image super-resolution, and image contrast enhancement respectively, and design fast and stable algorithms for the proposed models. The proposed image restoration method can be extended to many types of signal restoration problems and provide a novel idea and a new path for vision computing in the cloud era.