风格化图像通常采用艺术的表现手法，如素描、油画、水彩画等来描述场景，从而使得绘制效果能够更好地传达视觉焦点和情感。根据真实场景或照片重建风格化图像时，不仅要准确描绘场景内容，也要具备鲜明的艺术特征，达到风格化的目的。图像的风格化重建在图像编辑、动漫、娱乐甚至司法鉴证领域都有着广泛的应用，因此得到了越来越多研究者的关注。

• Image Transformation using Limited Reference with Application to Photo-Sketch Synthesis [Project]
  Wei Bai, Jiaying Liu, Yanghao Li and Zongming, IEEE Visual Communications and Image Processing (VCIP), Valletta, Malta, Dec. 2014. 
• Perspective Distorted Video Restoration and Stabilization for Mobile Devices (Under Review) [Project] 

     图像超分辨率指的是一种由一幅或多幅低分辨率图像预测其相应的高分辨率图像的技术，高分辨率意味着图像具有较高的像素分辨率，从而具有更多关键的细节和更清晰的图像质量。它在视频监控(Video surveillance)、图像打印 (Image printing)、刑侦分析 (Criminal investigation analysis)、医学图像处理 (Medical image processing)、卫星成像(Satellite imaging) 等领域有较广泛的应用。

• Nonlocal Based Super Resolution with Rotation Invariance and Search Window Relocation [Project]
  Yue Zhuo, Jiaying Liu, Jie Ren and Zongming Guo, IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Kyoto, Japan, Mar. 2012. 
• Segmentation-Based Scale-Invariant Nonlocal Means Super Resolution [Project]
  Saboya Yang, Jiaying Liu, Qiaochu Li and Zongming Guo, Asia Pacific Signal and Information Processing Association (APSIPA), Kaohsiung, Taiwan, Nov. 2013. 
• Context-Aware Sparse Decomposition for Image Denoising and Super-Resolution [Project]
  Jie Ren, Jiaying Liu, and Zongming Guo, IEEE Transaction on Image Processing (TIP), Vol.22, No.4, pp.1456-1469, Apr. 2013. IEEE International Conference on Pattern Recognition (ICPR), Tsukuba Science City, Japan, Nov. 2012.
• Image Super Resolution using Saliency-Modulated Context-Aware Sparse Decomposition [Project]
  Wei Bai, Saboya Yang, Jiaying Liu, Jie Ren and Zongming Guo, IEEE Visual Communications and Image Processing (VCIP), Sarawak, Malaysia, Nov. 2013.
• Sparse Representation Based Super Resolution Using Saliency and Edge Information [Project] 
  Saboya Yang, Jiaying Liu, Wenhan Yang and Zongming Guo, Asia Pacific Signal and Information Processing Association (APSIPA), Siem Reap, Cambodia, Dec. 2014.

     图像插值是根据图像已知采样点对未知点处数值进行预测(重采样)从而提高图像分辨率的过程。传统的插值算法如最近邻插值、双线性插值、双三次插值、Lanczos插值都是假设图像具有连续性，以此预测高分辨率点的像素值，这样容易导致相邻像素过于相似而产生锯齿或者模糊，因此，这些方法并不能解决图像放大过程中高频细节的问题。
     基于自回归模型的插值算法假设图像在局部区域内是稳定的，即图像在局部区域内的结构基本一致，于是可以用同一自回归系数来表征某一局部区域内的像素结构。分析自然图像的分段统计稳态性的基础上，可以利用基于概率描述的隐式分段自回归模型来刻画图像信号中的分段统计稳态性质，从而实现对图像信号较好的建模。此外，还进一步研究了图像的任意比例插值，扩展了图像插值的应用范畴。

• Adaptive General Scale Interpolation Based on Weighted Autoregressive Models [Project]
  Mading Li, Jiaying Liu, Jie Ren, Zongming Guo, IEEE International Symposium on Circuits and Systems (ISCAS), Beijing, China, May 2013. IEEE Trans. on Circuit System for Video Technology (TCSVT). Vol.25, No.2, pp.200-211, Feb. 2015.
• General Scale Interpolation Based on Fine-Grained Isophote Model With Consistency Constraint [Project]
  Wenhan Yang, Jiaying Liu, Mading Li and Zongming Guo 

     图像修复能够将图像中的缺失部分进行填充，从而提升图像的主观质量，达到让观察者看不出来“未曾被处理过”的效果。在去除椒盐噪声、错误隐藏、文字移除等领域有着广泛的应用。

• Image Restoration Based on 3-D Autoregressive Model via Low-Rank Minimization [Project]
  Mading Li, Jiaying Liu, Yu Guan and Zongming Guo, Data Compression Conference (DCC), Snowbird, Utah, Apr. 2015.
• Block-Based Multiscale Error Concealment Using Low-Rank Completion [Project]
  Mading Li, Jiaying Liu, Chong Ruan, Lu Liu and Zongming Guo, Asia Pacific Signal and Information Processing Association (APSIPA), Siem Reap, Cambodia, Dec. 2014

     基于图像块的离散余弦变换（Block-based discrete cosine transform，BDCT）被广泛应用于图像视频压缩中，然而，由于每个图像块在BDCT中是独立进行变换量化的，因此压缩后的图像在图像块之间经常会产生一些严重的失真，这种失真就是块效应。去块效应算法可以将压缩后在图像视频中产生的块效应去除，从而提升其主观质量。

• Patch-Based Image Deblocking Using Geodesic Distance Weighted Low-Rank Approximation [Project]
  Mading Li, Jiaying Liu, Jie Ren and Zongming Guo, IEEE Visual Communications and Image Processing (VCIP), Valletta, Malta, Dec. 2014.
• Image Blocking Artifacts Reduction via Patch Clustering and Low-Rank Minimization [Project]
  Jie Ren, Jiaying Liu, Mading Li, Wei Bai and Zongming Guo, Data Compression Conference (DCC), Snowbird, Utah, Mar. 2013. IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Vancouver, Canada, May 2013.