Scholarlink

Abstract <p>The work studies multichannel image visualization that is dimensionality reduction preserving perceptually critical information with inherent data loss. We investigate previously proposed Sokolov’s method, which leverages human perceptual differences between brightness and chromaticity. The assumption is the method yields maximally visually informative representation while minimizing perceptibility of visualization errors. To test this, we conducted a human study comparing Sokolov’s method against PCA and UMAP. Survey results from 62 subjects demonstrate Sokolov’s significant superiority in preserving local contrast.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Visualization of Multichannel Images Using Differences in Human Perception of Brightness and Chromaticity

  • 期刊论文

  • 发表日期：

    2026年7月20日

查看全文
  • D. S. Sidorchuk,
  • A. S. Bespalova,
  • A. A. Ukhanaeva

摘要

Abstract

The work studies multichannel image visualization that is dimensionality reduction preserving perceptually critical information with inherent data loss. We investigate previously proposed Sokolov’s method, which leverages human perceptual differences between brightness and chromaticity. The assumption is the method yields maximally visually informative representation while minimizing perceptibility of visualization errors. To test this, we conducted a human study comparing Sokolov’s method against PCA and UMAP. Survey results from 62 subjects demonstrate Sokolov’s significant superiority in preserving local contrast.