High-Fidelity Image Decolorization with Pure Odd Polynomial Subspace and Joint Covariance Optimization

Official Implementation for The Visual Computer

Manuscript Title: High-Fidelity Image Decolorization with Pure Odd Polynomial Subspace and Joint Covariance Optimization.

Overview: This repository provides the official source code, experimental data, and evaluation metrics for the decolorization frameworks (GPEP and COPD) described in the manuscript submitted to The Visual Computer. The goal is to enhance the transparency and reproducibility of our research on color-to-grayscale image conversion.

Repository Structure:

• src/: Contains the Python implementation of the Global Polynomial Eigen-Projection (GPEP) and Pure Odd Orthogonal Polynomials (COPD) algorithms.

• metrics/: Contains MATLAB scripts for calculating objective evaluation metrics, including CCPR, CCFR, E-score and Ec.

• dataset/: Includes sample images used for testing and demonstration.

• results/: Provides pre-generated grayscale results for verification.

• requirements.txt: Lists the necessary Python dependencies for running the code.

Affiliation: Developed by Lina Zhang, Jiale Yang, and Yamei Xu at Lanzhou University of Technology.

Citation: If you use this code or our research findings, please cite our manuscript published in The Visual Computer.