Scholarlink

<p>The color values of pixels in images and videos are always stored in a specific color profile. The light intensity captured by the camera sensor is transformed to a digital color vector according to a color transfer function of the given profile. This paper presents the results of an experimental comparison of the compression efficiency with different color profiles. State-of-the-art video and image codecs are used to compress 10 and 12-bit image sequences in different color profiles. The compression ratio and visual quality are measured after decompression and comparison with the original images. The results of this study suggest that the quality/size ratio depends on the selected profile and better results can be achieved, for example, with the AgX-Log profile. AgX-Log was identified as Pareto-optimal in 79% of the measurements and was second best in quality/size ratio in the rest. Its quality/size ratio is 1.63<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\times\)</EquationSource> </InlineEquation> higher compared to the standard sRGB profile (52.10 dB PSNR at 3.12 MB vs. 48.76 dB at 5.03 MB). The optimal transfer function exhibits a high-base logarithmic curve with gradual rise toward high intensities and range cropping. These findings were validated on three external datasets.</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

How color profile affects image and video compression efficiency

  • 期刊论文

  • 发表日期：

    2026年7月6日

查看全文
  • Tomáš Chlubna,
  • Pavel Zemčík

摘要

The color values of pixels in images and videos are always stored in a specific color profile. The light intensity captured by the camera sensor is transformed to a digital color vector according to a color transfer function of the given profile. This paper presents the results of an experimental comparison of the compression efficiency with different color profiles. State-of-the-art video and image codecs are used to compress 10 and 12-bit image sequences in different color profiles. The compression ratio and visual quality are measured after decompression and comparison with the original images. The results of this study suggest that the quality/size ratio depends on the selected profile and better results can be achieved, for example, with the AgX-Log profile. AgX-Log was identified as Pareto-optimal in 79% of the measurements and was second best in quality/size ratio in the rest. Its quality/size ratio is 1.63 \(\times\) higher compared to the standard sRGB profile (52.10 dB PSNR at 3.12 MB vs. 48.76 dB at 5.03 MB). The optimal transfer function exhibits a high-base logarithmic curve with gradual rise toward high intensities and range cropping. These findings were validated on three external datasets.