Research On The Edit Propagation For Certain Types Of Images

In Edit techniques for images, with the rise of digital media and software, has turned the way abundant and convenient in which people create and process images since it appeared in 1950s. And as one of the most classical research fields of edit technology, edit propagation receives much concern all the time. A number of methods have been developed practically; however, there are still lots of issues to be solved urgently.In this paper, new methods of edit propagation for three types of images are discussed. For high resolution image, a new down-sampling based Seam Carving is proposed to overcome the loss of details and regions concerned probably by users when down- sampling with classical methods. In our approaches, recalculation on probably important regions is used rather than up-sampling. For images with accompany, a pipeline is presented to detect accompany and realize associate process. Firstly, Hierarchical Segmentation for outline detection, and then FloodFill method notes target regions by seed flooding. SIFT algorithm search for the feature points of accompany finally. For texture images, referring to Fourier frequency spectrum method, Gabor filters is used to describe the texture. And stoke-preserved technique is introduced for colorization. On general framework for edit propagation, experiments with texture images show that method above extends the processing field of AppProp framework.Besides the analysis in mathematics and geometry, the implement of techniques above are introduced in detail. And the usability is evaluated with real high resolution images, images with accompany and texture images in related algorithms, which has showed strong availability of techniques discussed in this paper.