ComfyUI Node: Color Match

Color Match:

The ColorMatch node is designed to facilitate the transfer of color characteristics from one image to another, making it an invaluable tool for tasks such as automatic color grading of photographs, paintings, and film sequences, as well as light-field and stop-motion corrections. By leveraging advanced color transfer methods, including those proposed by Reinhard et al., Pitie et al., and a Multi-Variate Gaussian Distribution (MVGD) transfer combined with classical histogram matching, this node ensures high-quality and consistent color matching. The node's primary goal is to enable seamless and efficient color harmonization across different images, enhancing the visual coherence and aesthetic appeal of your projects.

Color Match Input Parameters:

image_ref

This parameter represents the reference image whose color characteristics will be transferred to the target image. The reference image serves as the source of the desired color palette. It is crucial to ensure that the reference image is either a single image or a batch of images that match the batch size of the target images. This parameter is required and must be of type IMAGE.

image_target

This parameter represents the target image to which the color characteristics from the reference image will be applied. The target image is the one that will undergo color transformation to match the reference image. Similar to the reference image, the target image can be a single image or a batch of images. This parameter is required and must be of type IMAGE.

method

This parameter specifies the method used for color transfer. The available options are mkl, hm, reinhard, mvgd, hm-mvgd-hm, and hm-mkl-hm. Each method employs a different algorithm for color matching, offering various levels of accuracy and visual results. The default method is mkl. Selecting the appropriate method can significantly impact the quality and style of the color transfer.

Color Match Output Parameters:

image

This output parameter represents the resulting image after the color transfer process. The output image will have the color characteristics of the reference image applied to the target image, resulting in a visually coherent and color-matched image. The output is of type IMAGE and is clamped to ensure pixel values remain within the valid range of 0 to 1.

Color Match Usage Tips:

• To achieve the best results, ensure that the reference and target images have similar content and lighting conditions. This will help the color transfer process produce more natural and visually appealing results.
• Experiment with different color transfer methods to find the one that best suits your specific project needs. Each method has its strengths and may produce different visual outcomes.
• When working with batches of images, ensure that the batch sizes of the reference and target images match to avoid errors and ensure consistent color transfer across all images in the batch.

Color Match Common Errors and Solutions:

Can't import color-matcher, did you install requirements.txt? Manual install: pip install color-matcher

• Explanation: This error occurs when the required color-matcher library is not installed.
• Solution: Install the color-matcher library by running the command pip install color-matcher in your terminal.

Color Match: Use either single reference image or a matching batch of reference images.

• Explanation: This error occurs when the batch size of the reference image does not match the batch size of the target image.
• Solution: Ensure that the reference image is either a single image or a batch of images that match the batch size of the target images.

Error occurred during transfer: <specific_error_message>

• Explanation: This error occurs when there is an issue during the color transfer process, such as incompatible image dimensions or data types.
• Solution: Check the specific error message for details and ensure that the input images are compatible in terms of dimensions and data types. Make any necessary adjustments to the images and try again.