About convert jpg to nrrd

Dear PyRadiomics / 3DSlicer community,
When importing JPEG images into 3D Slicer and converting them to the NRRD format, they may not be recognized by the Radiomics module. However, by converting them to a scalar volume, they can be recognized by the Radiomics module, although the image becomes grayscale. Is there a way to convert the images to a scalar volume while preserving the color information, or are there other methods to make them recognizable by the Radiomics module?
Could you please guide me on how to accomplish this task.
Thank you very much.

There may not be any metric of PyRadiomics that is computed from multichannel images (such as RGB color images). You can use one color channel, a combined channel (e.g., average), or compute features for each color channel and combine the features.