Abstract
We consider medical image transformation problems where a grayscale image is transformed into a color image. The colorized medical image should have the same features as the input image because extra synthesized features can increase the possibility of diagnostic errors. In this paper, to secure colorized medical images and improve the quality of synthesized images, as well as to leverage unpaired training image data, a colorization network is proposed based on the cycle generative adversarial network (CycleGAN) model, combining a perceptual loss function and a total variation (TV) loss function. Visual comparisons and experimental indicators from the NRMSE, PSNR, and SSIM metrics are used to evaluate the performance of the proposed method. The experimental results show that GAN-based style conversion can be applied to colorization of medical images. As well, the introduction of perceptual loss and TV loss can improve the quality of images produced as a result of colorization better than the result generated by only using the CycleGAN model.
Highlights
Compared to grayscale images, color images contain detail and clear information so that, in many applications, when only grayscale images are generated, the grayscale image is often transformed into a color image first
& We demonstrate that the introduction of perceptual loss and total variation (TV) loss can improve the quality of images produced as a result of colorization, and the result is better than the result generated by only using the CycleGAN model
The experiment results show that the Perceptual method has better performance than CycleGAN
Summary
Color images contain detail and clear information so that, in many applications, when only grayscale images are generated, the grayscale image is often transformed into a color image first. Many studies have been done in various fields to automatically and effectively generate color images [1]. Multimedia Tools and Applications takes advantage of the well-established radar polarimetry theories and polarimetry approaches to interpret the vast majority of SAR images that are not full-pol. A semi-automatic colorization system takes a monochrome manga and reference images as input and generates a plausible color version of the manga [8]. Most medical image datasets are photographed as grayscale images, so it is difficult to obtain enough color images. To solve this problem, a colorization operation for converting monochrome images into color images should be considered
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
