Color quantization

Color quantization refers to the process of reducing the number of colors in a digital image by quantizing the color gamut of the image, resulting in a lossy image compression operation. This technique is used to improve computational efficiency and reduce the histogram size of the image. Color quantization information is applied to each pixel of the image. There are two main methods of color quantization: clustering and segmentation.From: Digital Color Imaging Handbook [2017], Understanding Digital Image Processing [2018], Computer Graphics [2019], Recognition and analysis of fabric texture by double-sided fusion of transmission and reflection images under compound light source [2022]

Image Retrieval

View Chapter

Purchase Book

Published in Vipin Tyagi, Understanding Digital Image Processing, 2018

Vipin Tyagi

The most commonly-used color descriptor is a color histogram. It extracts the global distribution of color among the pixels. However, spatial relation among pixels of the same color is ignored by this descriptor. Color auto-correlogram relates the color information of the pixels of the same color at the specified distance. Further, the use of color quantization is embedded in order to provide better computational efficiency and reduced histogram size. Color coherent vector (CCV), color correlogram, color moment, joint histogram, color histogram, color autocorrelogram, MPEG-7 color descriptor and dominant color descriptor are the other descriptors used for describing the color features of the image.

Recognition and analysis of fabric texture by double-sided fusion of transmission and reflection images under compound light source

View Article

Journal Information

Published in The Journal of The Textile Institute, 2022

Mingzhu Fan, Na Deng, Binjie Xin, Runhu Zhu

There are two main methods of color quantization: clustering and segmentation. In this paper, the color quantization problem of RGB weave float image is transformed into pixel clustering problem by using the color quantization algorithm based on clustering. The mean square error of the final clustering result is shown in formula (15).

Color quantization

Explore chapters and articles related to this topic

Image Retrieval

Recognition and analysis of fabric texture by double-sided fusion of transmission and reflection images under compound light source