The importance of the normalizing channel in log-chromaticity space

Abstract

The log-chromaticity space (LCS) is a color space with excellent illumination-invariant properties. When converting RGB colors to LCS, there exist four different options for choosing the normalizing channel. For classification applications, we analyze the impact of the normalizer on the distribution of colors in LCS. Based on synthetic and real image data we show that the geometric mean does not introduce a bias to the color clusters and always results in an intermediate clustering performance. However, data-specific selection of the normalizing channel can further improve the results. For instance, for skin classification we show that using the blue channel as denominator results in a recognition improvement of about 25.9% compared to the red channel (worst result).In comparison to the geometric mean and the green channel, the two most popular denominators, the performance increase is 12.9% and 2.9%.

CF-StyleGAN: Near-infrared image colorization of SE attention StyleGAN via color features

“…As a result, the LAB color space is conducive to improving NIR image colorization accuracy. In reference to color constancy literature [16], we utilize histogram features to minimize the effect of brightness variations, and we construct the histogram as a differentiable histogram of colors in logarithmic chromaticity space [17]. Here, the feature is a 2D histogram of the image colors projected into the log-chromatic space.…”

HistoGAN: Controlling Colors of GAN-Generated and Real Images via Color Histograms

“…The histogram feature used by HistoGAN is borrowed from the color constancy literature [4,6,10] and is constructed to be a differentiable histogram of colors in the log-chroma space due to better invariance to illumination changes [19,21]. The feature is a 2D histogram of an image's colors projected into a log-chroma space.…”