Shades of Gray and Colour Constancy

By Graham David Finlayson, E. Trezzi. International Conference on Communications in Computing, 2004

721 citations recorded by the source corpus.

Colour constancy is a central problem for any visual system performing a task which requires stable perception of the colour world. To solve the colour constancy problem we estimate the colour of the prevailing light and then, at the second stage, remove it. Two of the most commonly used simple techniques for estimating the colour of the light are the Grey-World and Max-RGB algorithms. In this paper we begin by observing that this two colour constancy computations will respectively return the right answer if the average scene colour is grey or the maximum is white (and conversely, the degree of failure is proportional to the extent that these assumptions hold). We go on to ask the following question: “ Would we perform better colour constancy by assuming the scene average is some shade of grey?”. We give a mathematical answer to this question. Firstly, we show that Max-RGB and Grey-World are two instantia-tions of Minkowski norm. Secondly, that for a large cali-brated dataset L6 norm colour constancy works best over-all (we have improved the performance achieved by a sim-ple normalization based approach). Surprisingly we found performance to be similar to more elaborated algorithm.

Subfields: Atomic and Molecular Physics, and Optics, Cell Biology, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/acf36cdadfec869f136602ea41cad8b07e3f8ddb

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Atomic and Molecular Physics, and Optics

In Cell Biology

In Computer Vision and Pattern Recognition

Citations in the corpus

Learning Image-Adaptive 3D Lookup Tables for High Performance Photo Enhancement in Real-Time (2020, background): “For the enhancement part, those modules include exposure correction [13], [14], color constancy [15], [16], contrast enhancement [17], [18], color manipulation [19], [20] and tone mapping [21], [22], each of which has been widely studied in the past.”
Referenceless Prediction of Perceptual Fog Density and Perceptual Image Defogging (2015, methodology): “The shades-of-gray color constancy technique [52] is used because it was fast and”
Color Balance and Fusion for Underwater Image Enhancement (2018, methodology): “On this set of images, we applied the following methods: the classical white-patch max RGB algorithm [42], the Gray-World [43], but also the more recent Shades-of-Grey [41] and Gray-Edge [40].”
Emerging From Water: Underwater Image Color Correction Based on Weakly Supervised Color Transfer (2017, background): “For underwater image color correction, traditional color constancy methods ( e.g. , Gray World [20], Max RGB [21], White Balance [22], Shades-of-Grey [23], and et al. ) and their variations are usually employed.”
Data-Driven Color Augmentation Techniques for Deep Skin Image Analysis (2017, methodology): “In this work, we estimate the illuminant on images from the training set applying the Shades of Gray color constancy technique [11], as implemented by the general color constancy framework described in [12].”
Bayesian color constancy revisited (2008, background): “Each of the approaches models image intensity as a product of a uniform illuminant with a reflectance function over the visible scene.”
Computational Color Constancy: Survey and Experiments (2011, positive): “In [60], the white patch and the gray-world algorithms are shown to be special instantiations of the more general Minkowski framework”
Edge-Based Color Constancy (2007, positive): “We further extend the color constancy framework of Finlayson and Trezzi [15] to also include color constancy methods derived from the Grey-Edge hypothesis.”
Single Image Dehazing by Multi-Scale Fusion (2013, methodology): “Despite of its simplicity, this lowlevel approach of Finlayson and Trezzi [37] has shown to yield comparable results to those of more complex white balance algorithms (that produces reliable results based on natural image statistics [40]).”
STAR: A Structure and Texture Aware Retinex Model (2019, neutral): “Method White-Patch [13] Gray-Edge [56] Shades-Gray [55] MAE↓ 7.”
Enhancing underwater images and videos by fusion (2012, methodology): “The Finlayson’s approach Shades-of-Grey [12] computes the illumination of the scene for each channel by using the Minkowski p -norm.”