2 software implementations in PyTorch and Matlab for the CIE XYZ Net image unprocessing framework. This codebase enables the transformation of processed images back to the device-independent CIE XYZ color space as described in the TPAMI 2021 paper.
Use Cases
- Transform processed images into the CIE XYZ space using the provided PyTorch model.
- Implement image unprocessing stages in Matlab-based vision pipelines.
- Reproduce the unprocessing benchmarks described in the TPAMI 2021 paper.
Strengths
- Includes source code for both PyTorch and Matlab implementations.
- Implements the architecture from the TPAMI 2021 paper 'CIE XYZ Net: Unprocessing Images for Low-Level Computer Vision Tasks'.
- Provides a neural network architecture specifically for mapping images to the CIE XYZ color space.