Investigation of texture effect on visual colour difference evaluation

John H. Xin, Hui‐Liang Shen, Thomas Chuen LamColor 2005 · DOI

TL;DR

This paper quantifies how fabric texture influences our perception of color differences. They found that texture significantly impacts how we judge color variations, with higher texture strength leading to a perceived decrease in color difference. This has implications for digital color communication and quality control in industries like textiles.

What Problem Does This Paper Address?

In fields like textiles, product design, and quality control, accurately communicating and evaluating color is crucial. While electronic displays like CRTs are widely used, they often present color samples with inherent textures (like woven fabrics). The problem is that the visual perception of color difference can be significantly altered by these textures, a phenomenon known as the "texture effect." While this effect is acknowledged, there's a lack of quantitative analysis on how different texture levels influence color difference evaluation. Existing studies have been limited in scope, making it difficult to incorporate this effect into color difference models or digital workflows. This paper aims to provide a quantitative understanding of this texture effect, specifically focusing on how varying texture levels affect the perceived color difference.

Methods

The researchers developed a method to generate textured color samples for evaluation. They started with 15 scanned textile fabric images, representing various woven patterns. To quantify texture, they used the "half-width of histogram" of the luminance channel, termed "texture strength." This metric captures the coarseness of the texture.

They then employed a "color-mapping technique" to apply specific color differences to these textured images. This involved taking a base color and modifying its luminance channel pixel-by-pixel based on the texture's luminance distribution. This allowed them to create textured color pairs with a controlled color difference of approximately 5.0 CIELAB units in the lightness direction.

For visual evaluation, they used a calibrated CRT monitor in a dark room. Observers were presented with these textured color pairs alongside a gray scale. They were asked to assign a gray scale grade that best matched the perceived color difference of the textured pair. This "gray-scale comparison method" is a standard in the textile industry. They also conducted a reference experiment with solid color pairs to compare against.

Datasets

The study utilized 15 scanned textile fabric images, representing common woven patterns like plain, twill, rib, jacquard, and basket weaves. These images were used to generate textured color samples.

The colorimetric values for five specific "color centers" (Orange, Yellow, Gray, Green, Blue) were selected based on previous research indicating regions where different color difference formulas show disagreement. These color centers were then used as the base for generating the textured color pairs.

A gray scale with five grades was used as a reference for visual assessment.

Key Insights and Ideas

  • Texture is a significant parametric factor: The study strongly confirms that fabric texture is not just a visual background but actively influences how we perceive color differences.
  • Texture strength matters: The "texture strength," quantified by the histogram half-width, directly correlates with the perceived color difference.
  • Perceived difference decreases with texture: Counterintuitively, as the texture becomes stronger (higher texture strength), the perceived color difference between textured pairs tends to decrease. This suggests that texture can mask or alter our sensitivity to color variations.
  • Observer agreement improves with texture: Textured pairs showed less variation in observer judgments (lower standard deviation) compared to solid color pairs, indicating that texture can lead to more consistent evaluations.
  • Luminance is key: The study focused on luminance differences, aligning with the understanding that the human visual system is more sensitive to luminance variations than chromatic ones.

Key Results

  • A high correlation was found between texture strength and visual color difference for textured pairs.
  • An increase of 10 units in texture strength (in luminance scale) was associated with a decrease of approximately 0.25 units in perceived visual difference.
  • The ratio of visual difference between textured and solid color pairs (Kw) indicated a strong parametric effect of texture. When texture strength was low, Kw was closer to 1.0, but it decreased as texture strength increased, signifying a stronger texture effect.
  • The standard deviation of visual differences for solid color pairs was larger than for textured pairs, suggesting observers were more consistent in judging textured samples.

Limitations and Future Work

  • Texture Representation: The study primarily used the histogram half-width (texture strength) to represent texture. The authors acknowledge that this metric might not capture all textural nuances, as two different textures could have similar histograms. They suggest future work could incorporate additional textural features like coarseness, contrast, and complexity.
  • Color Difference Magnitudes: This study focused on a medium color difference of approximately 5.0 CIELAB units. Future research could investigate the texture effect on smaller color differences, which are critical for industrial quality control.
  • Hue and Chroma: While the study focused on lightness, the authors note that hue and chroma might also be influenced by texture. Further investigation into these aspects is recommended.
  • Broader Investigation: The authors suggest that more color differences need to be investigated to fully integrate the parametric effect of texture into color difference equations.

Summary updated

Abstract

Abstract The texture effect on visual colour difference evaluation was investigated in this study. Five colour centers were selected and textured colour pairs were generated using scanned textile woven fabrics and colour‐mapping technique. The textured and solid colour pairs were then displayed on a characterized cathode ray tube (CRT) monitor for colour difference evaluation. The colour difference values for the pairs with texture patterns are equal to 5.0 CIELAB units in lightness direction. The texture level was represented by the half‐width of histogram, which is called texture strength in this study. High correlation was found between texture strength and visual colour difference for textured colour pairs, which indicates that an increasing of 10 units of texture strength in luminance would cause a decreasing of 0.25 units visual difference for the five colour centers. The ratio of visual difference between textured and solid colour pairs also indicates a high parametric effect of texture. © 2005 Wiley Periodicals, Inc. Col Res Appl, 30, 341–347, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.
BibTeX
@inproceedings{xin2005investigation,
  author = {John H. Xin and Hui‐Liang Shen and Thomas Chuen Lam},
  title = {{Investigation of texture effect on visual colour difference evaluation}},
  booktitle = {Color},
  year = {2005},
  doi = {10.1002/col.20138},
}

Loading related content...

Figure