opencv
Integrate ICC v5 / iccMAX Support into OpenCV
Dear OpenCV Maintainers,
I am writing to propose an enhancement to OpenCV’s color management capabilities: the integration of ICC v5 / iccMAX profile support. As imaging workflows increasingly rely on wide-gamut, high-dynamic-range content, modern color management standards are essential for accurate processing, visualization, and output.
Motivation Current OpenCV workflows typically rely on ICC v2/v4 profiles and manual gamma approximations for wide-gamut color spaces (e.g., ProPhoto RGB, ROMM RGB) or HDR content. This approach has several limitations:
Limited channel support: Only 3-channel RGB is fully supported; multi-channel or spectral data cannot be handled.
Approximate gamma and clipping: Wide-gamut transforms require manual gamma adjustments, introducing artifacts and loss of fidelity.
No perceptual color appearance modeling: Viewing conditions, adaptation, and flare corrections are not considered.
Limited HDR support: Floating-point and extended dynamic range are only partially implemented, restricting true HDR workflows.
Proposed Upgrade Integrate ICC v5 / iccMAX profile handling directly into OpenCV, leveraging:
Multi-channel LUTs and transforms: Accurate conversion between wide-gamut color spaces and high-bit-depth or spectral data.
High-precision floating-point processing: Maintain HDR, linear-light, and extended dynamic range data throughout pipelines.
Color appearance models (CAM02 / CAM16): Support perceptual transforms for soft-proofing and viewing condition adaptation.
Backward compatibility: Ensure existing workflows using v2/v4 profiles remain functional.
Benefits
Professional color accuracy: OpenCV would handle modern cinema cameras (RED, Arri, Blackmagic) and high-end photography workflows accurately.
HDR and wide-gamut support: True representation of HDR content without manual approximations.
Scientific and industrial applications: Enhanced precision for research, medical imaging, and machine vision systems.
Enhanced Python and C++ pipelines: Enable real-time processing of color-managed content for computer vision, machine learning, and display.
Improved dataset quality: Training datasets generated in OpenCV would more accurately reflect real-world color and lighting, improving model generalization.
Suggested Implementation
Integrate an ICC v5 parser and transform engine (e.g., leverage DemoIccMax / iccDEV).
Provide Python bindings for profile application in cv2.cvtColor() or a new cv2.colorProfileTransform() function.
Maintain internal floating-point representation for linear color processing.
Add support for perceptual adaptation and multi-channel LUTs.
Adding ICC v5 / iccMAX support would bring OpenCV to parity with professional imaging software and ensure Python, C++, and GPU-based pipelines can process true wide-gamut HDR content accurately. This upgrade would significantly benefit research, industrial, and machine vision applications, positioning OpenCV as a robust choice for modern imaging workflows. Integrating the iccMAX ICC file by default
Thank you for considering this enhancement. https://www.color.org/iccmax.xalter
