Region Filling with Super Resolution Algorithm

Region filling which has another name inpainting, is an approach to find the values of missing pixels from data available in the remaining portion of the image. The missing information must be recalculated in a distinctly convincing manner, such that, image look seamless. This research work has built a methodology for completely automating patch priority based region filling process. To reduce the computational time, low resolution image is constructed from input image. Based on texel of an image, patch size is determined. Several low resolution image with missing region filled is generated using region filling algorithm. Pixel information from these low resolution images is consolidated to produce single low resolution region filled image. Finally, super resolution algorithm is applied to enhance the quality of image and regain all specifics of image. This methodology of identifying patch size based on input fed has an advantage over filling algorithms which in true sense automate the process of region filling, to deal with sensitivity in region filling, algorithm different parameter settings are used and functioning with coarse version of image will notably reduce the computational time.