Overview of Grayscale Image Colorization Techniques

Abstract

Colorization is the process of introducing color to grayscale digital images. In these images, each pixel has a scalar value representing its intensity. However, the pixels of color images contain more complex, three-dimensional information. Depending on a color model, the pixel attributes correspond to a three-value color representation. For the purpose of grayscale image colorization, we strongly advise the use of color representation, which allows easy luminance extraction. This makes the colorization simpler-we only need to estimate two chrominance channels, while the luminance of the colorized image remains unchanged.Despite plenty of color models, we cannot directly transform a one-dimensional grayscale value to a three-dimensional color space. We only get the luminance of the colorized image, while the two remaining color channels are still unknown. Therefore, the colorization may be performed in many different ways, achieving various outcomes. Additionally, it is difficult to objectively assess the final colorization results. We can compare the colorized version using standard image quality measures only if we have the color version of the image. The colorization of grayscale images without a reference has to be evaluated mainly subjectively.In general, there are two main types of colorization frameworks. The first one is the semiautomatic approach. In such a range of methods, a user has to indicate the color hints in form of scribbles within the grayscale image. The scribbles enable the algorithms deciding which colors have to be put in corresponding parts of the image. The colorization procedure in the semiautomatic approach may be visualized as color spilling over the image areas starting from the indicated scribbles.The automatic approach employs a color image that resembles the colorized grayscale picture. The algorithms rely on the similarities between the luminance B. Smolka ( ) · A. Popowicz

Medical Image Colorization for Better Visualization and Segmentation

“…Colorized medical images assist in prevention and treatment of diseases. Colorization techniques can be classified into three types -automatic, semi-automatic and user defined coloring techniques [21].…”

Artifacts Reduction Using Multi-Scale Feature Attention Network in Compressed Medical Images

“…Deep learning methods using Convolutional Neural Network (CNN) have brought great potentials into low-level computer vision applications such as Super Resolution (SR) [2][3][4][5][6][7][8], image denoising [9][10][11][12][13][14][15][16], and image colorization [17,18]. In particular, these applications have been developed by CNN-based image denoising methods with deeper and denser network architectures [19,20].…”

A Proposed HSV-Based Pseudo Coloring Scheme for Enhancing Medical Image

“…When taking into consideration decorrelated color spaces, where the channels are orthogonal, we may assess the quality with the root of the sum of squared SSIM results obtained for each channel. The detailed Mean SSIM (MSSIM) could be found in [15].…”

“…SSIM reflects how the colorization process affects the structure of the image. Popowicz [15] improved the SSIM by adding a color comparison to the criteria of the grayscale SSIM. Since the SSIM is defined only for grayscale images, it can be adapted for color image colorization by calculating the SSIM for every single color channel independently, then calculating the mean.…”

“…From the table, it seems the proposed palettes have high MSE and low PSNR. Since NCD is the most suitable measure for colorization procedures assessment [15] it is considered for assessing the 12 methods. The lower the NCD, the better the image quality.…”