Categorizing natural color distributions

Abstract

The natural objects that we are surrounded with virtually always contain many different shades of color, yet the visual system usually categorizes them into a single color category. We examined various image statistics and their role in categorizing the color of leaves. Our subjects categorized photographs of autumn leaves and versions that were manipulated, including: randomly repositioned pixels, leaves uniformly colored with their mean color, leaves that were made by reflecting the original leaves' chromaticity distribution about their mean ("flipped leaves"), and simple patches colored with the mean colors of the original leaves. We trained a linear classifier with a set of image statistics in order to predict the category that each object was assigned to. Our results show that the mean hue of an object is highly predictive of the natural object's color category (>90% accuracy) and observers' choices are consistent with their use of unique yellow as a decision boundary for classification. The flipped leaves produced consistent changes in color categorization that are possibly explained by an interaction between the color distributions and the texture of the leaves.

Development of a machine learning model for extracting image prominent colors

“…As outlined in the Introduction, some studies suggest that the perception of an object's color category highly depends on the object's mean hue. 8 For that reason, we have defined a new metric that takes into account the distribution of swatch's hue. Let H(x, y) be the hue and C (x, y) the chroma channel of the image in CIELCh color space.…”

“…Since the HVS detects object's colors as a single color category, 8 it is reasonable to assume that image segments are essential for the perception of prominent colors. For this reason, we used a graph-based method presented by Felzenszwalb and Huttenlocher 45 to segment the images and analyze the recoloring error by segments.…”

“…4 Although natural objects contain different shades of color, our visual system usually detects them as a single color category. 8 By investigating single color impression of multicolored texture patterns or natural objects, several researchers came to the conclusion that color impression cannot be simply determined by a colorimetric average of elemental colors but is influenced by a mechanism that integrates the color appearance attributes (hue, saturation, and brightness) of elemental colors. [9][10][11] Moreover, some authors claim that an overall color impression strongly depends on the most saturated elemental colors, which result in a chromatic enhancement in perceived single color impression.…”

“…For example, Milojevic et al discovered that the natural object's color category largely depends on the object's mean hue and is not shifted toward the most saturated colors. 8 The aforementioned inconsistencies in results could be contributed to different methodological approaches-color matching vs color categorization.…”

The Inner Image of Russian Color Term черньӀй (Black) (Experimental Study)

“…Nowadays the group of color terms (CT) is one of the most intriguing and promising spheres for a linguistic study. "The natural objects that we are surrounded with virtually always contain many different shades of color, yet the visual system usually categorizes them into a single color category" [1]. Of particular interest is the question of how our knowledge of color as physical and mental phenomenon is related to our knowledge of color naming.…”