EPIIC: Edge-Preserving Method Increasing nucleI Clarity for Compression Artifacts Removal in Whole-Slide Histopathological Images

This repository contains the implementation of the EPIIC and EPIIC Sobel methods designed to remove JPEG compression artifacts from hematoxylin and eosin (H&E)-stained histopathological images.

How to use the code

1. Start by running the main.m script, or add all folders in the repository to your MATLAB path.

2. Then, navigate to the test_algorithm.m script, where you can run the methods on example images from the BreCaHAD dataset (located in the Data folder).

3. You can customize the algorithm parameters:

   • Q – JPEG compression quality factor
   • Sigma – standard deviation of the Gaussian filter
   • T_mult – T_m parameter (see details in the manuscript)
   • avg_size – size of the averaging filter
   • he – whether the image is H&E-stained (true enables stain deconvolution)
   • save – whether to save intermediate result maps at each step (true/false)
   • method – choose "EPIIC" or "EPIIC Sobel"
     • Other tested methods are available in the Utils/process_image.m function

4. Run the code, and the result with SSIM and PSNR values will be displayed in the Figure window.

How to cite

If you use this repository in your research, please cite the following paper:

Merta, J., & Marczyk, M. (2025). EPIIC: Edge-Preserving Method Increasing Nuclei Clarity for Compression Artifacts Removal in Whole-Slide Histopathological Images. Applied Sciences, 15(8), 4450.