A robust adaptive super-resolution method

摘要：

To get robust results under varied noise, an adaptive super-resolution (SR) method is proposed. The method runs in frame of regularization, and there will be an iteration process, the signal to noise ratio(SNR) is estimated by the local varicance of image in every iteration step, and be used as the regularization parameter to control the iteration, as the SNR is a reflection of image noise, and varied with iteration, the iteration is adaptive and converges to a optimal solution. The results show that the proposed method is effective, and has robust results under varied noise.

关键词： super-resolution regularization SNR noise

作者: Nan Li Xiaobo Lu

作者单位: School of Transportation Southeast University Nanjing, Jiangsu, China School of Automation Southeast University Nanjing, Jiangsu, China

会议类型: 国际会议

会议名称: 2011 Third International Conference on Intelligent Human-Machine Systems and Cybernetics 第三届智能人机系统与控制论国际会议 IHMSC 2011

会议地点: 杭州

会议语种:英文

页码: 314-317

在线出版日期: 2011-08-26（万方平台首次上网日期，不代表论文的发表时间）