Image completion with variable scope patch sampling
Jing Liu, Qing Zhu · 2016
This paper presents a simple but effective way of patch sampling for image completion which changes the assignment scope in different pixels when sampling patches during completion. In each pixel of missing regions, the scope expands from narrow to wide and stop expanding as soon as it randomly find a patch in searching region. In this way, all the sampled patches are almost around the missing regions and each pixel in the regions is assigned an offset value which is corresponds to one of the patches. Traditional methods simply assign the offsets from the whole images and then use algorithms (ANN, etc.) for further process. Our method reduces misleading results and is easier to acquire better patches than many other patch-based methods. Results show that with the same input images, the proposed method obtains better results than others.