Coupled Total Variation And Its Application In Image Colorization

Image colorization technique is an active and challenging subject in the field of digital image processing.At present,it has been widely used in the medical treatment,film-television industry,space exploration and video surveillance and so on.In particular,it is undoubtedly helpful for clinicians to observe and diagnose diseases by coloring such grayscale medical images as US,CT,MRI images and so on.In this paper,we focus on the image colorization method based on coupled total variation.The main research contents and innovations are as follows:(1)Firstly,several image colorization models based on variations partial differential equations are introduced,such as Sapiro's color inpaint model,Kang's model and improved Kang's model.Then the advantages and disadvantages of the above models are simply analyzed.In addition,we give three fast algorithms for numerically solving the TV model,including dual algorithm,Split Bregman algorithm and ADMM algorithm.(2)On the basis of the existing colorization models,we propose a coupled TV model for image colorization by combining the YCbCr color space and Prewitt edge detection operator.Then by incorporating the ADMM method,we design a fast numerical algorithm for numerically solving the proposed model,and give the convergence of the algorithm.Numerical experiments results are reported to demonstrate that the proposed model can effectively overcome the problem of edge crossover of colors while colorizing large regions in grayscale images well.(3)The existing methods of the recognition and feature extraction for brain tumors in MRI images are detailed introduced.Then we colorize the MRI images based on brain tumors by using the proposed coupled TV model.Experimental results show that the proposed model can effectively recognize the brain tumor regions by comparing some existing segmentation results.