Image compression and restoration using inpainting algorithms for wireless communication applications

Abstract

quotDigital image inpainting is a technique of restoring large removed /dam aged regions in an image with the data from the surrounding pixels ofthe removed region. The problem of image restoration with inpainting techniques occurs commonly in image computer vision/image processing when unwanted objects have to be removed from images, or for filling cracks in photographs etc. Recently, Digital Image Inpainting approach is an active field of research in two significant applications of wireless com munication such as image compression and recovering the image from the damaged image due to errors in wireless channel. The earlier works in newlineliterature emphasize on sending assistant parameters for more reliable im age compression using Inpainting algorithms and for channel impairments on image, error correction strategies for the reliable information recoverywith redundant bits which in turn decreases bandwidth efficiency of the transmission. newlineThe contribution of this thesis is to develop inpainting algorithms taking into consideration of the above two scenarios. To achieve higher compres sion ratios, with less bandwidth, hybrid image compression algorithms are developed by combining inpainting algorithm with state-of-the-art com pression standards such as JPEG. The set goal is to obtain a compression ratio up to 1:20 with PSNR 25 dB to 40 dB using inpainting algorithms. newlineTo restore the image from the channel effects, image propagation behavior in fading channel (Rayleigh / Rice) is studied and inpainting techniques are applied to restore the image. The set goal is to restore the quality image in fading environment with PSNR 25 dB to 35 dB, without adding redundant bits, so as to increase bandwidth efficiency and reduce the processing time. newlineThe research work results prove that at plausible levels, the PSNR of 28 dB to 34 dB, a compression ratio 4.87 - 17.62 (1.64bpp - 0.45bpp) are obtained, resulting in 80% to 94.33% bits-savings.