摘要
简要介绍了图像修补问题及现在一些主要的图像修补技术。根据图像修补的目的,提出小波图像修补问题不必局限于小波域求解,也可以在空域得到解决的思想。给出了该思想的依据和具体实现步骤,并提出一种在空域中基于Rudin-Osher-Fatemi(ROF)模型来实现上述思想的小波图像修补方法。图像修补仿真实验的结果表明,该方法可以得到较好的图像修补视觉效果,峰值信噪比(PSNR)小于10dB的缺损图像经处理后其PSNR达到约30dB。较以前的小波图像修补方法,该方法显著提高了缺损图像的PSNR,验证了提出思想的可行性。
Image inpainting and some of its main techniques are briefly presented.To inpaint image,an idea to solve wavelet image inpainting extending from wavelet domain to pixel domain and its theory and implement processing are given.A wavelet image inpainting method in pixel domain based on Rudin-Osher-Fatemi（ROF）model is proposed to put the idea above into practice.Experimental results show that the method can get good visual effect of image inpainting,with higher peak signal to noise ratio（PSNR）compared with former methods,and the PSNR of damaged image increases from less than 10 dB to 30 dB after processing using the proposed method.Therefore,the wavelet image inpainting is available in pixel domain,which means that the proposed method is practicable.
出处
《激光与光电子学进展》
CSCD
北大核心
2010年第10期71-75,共5页
Laser & Optoelectronics Progress
基金
国家自然科学基金(60872138)资助课题
关键词
图像处理
图像修补
ROF模型
小波
image processing
image inpainting
Rudin-Osher-Fatemi model
wavelet
