An Automatic Image Inpainting Algorithm Based on FCM

There are many existing image inpainting algorithms in which the repaired area should be manually determined by users. Aiming at this drawback of the traditional image inpainting algorithms, this paper proposes an automatic image inpainting algorithm which automatically identifies the repaired area by fuzzy C-mean (FCM) algorithm. FCM algorithm classifies the image pixels into a number of categories according to the similarity principle, making the similar pixels clustering into the same category as possible. According to the provided gray value of the pixels to be inpainted, we calculate the category whose distance is the nearest to the inpainting area and this category is to be inpainting area, and then the inpainting area is restored by the TV model to realize image automatic inpainting.