Image inpainting based on particle swarm ptimization

Abstract EN

Digital image in painting is a technique that can repair a portion of damaged or removed image by means of automatic mechanisms.

This paper proposes an automatic approach for reconstruction of small missing or damaged portions of images using Particle Swarm Optimization (PSO).

The PSO inpainting approach attempts to find the best match between Red, Green, and Blue (RGB) channels of the pixels masked (i.e.

pixels of the region to be removed) and pixels in the surrounding area, and transfer the color of the best matching.

The results show that our mechanism is more efficient (i.e.

less time required to inpaint the masked region) when compared with genetic-based image inpainting.