WSI-APIC

This repository contains the codes and demonstration data for a Whole Slide Imaging system based on Angular Ptychographic Imaging with a Closed-form Solution (WSI-APIC). It includes GPU-accelerated APIC reconstruction algorithms and sample location segmentation codes.

Paper link: https://doi.org/10.1364/OPTICA.505283

arXiv: https://arxiv.org/abs/2310.18529

Original APIC paper link: https://doi.org/10.1038/s41467-024-49126-y

Original APIC code repository: https://github.com/rzcao/APIC-analytical-complex-field-reconstruction

Top-level folder structure:

├── Data                              # Directory containing raw APIC data and calibration data for preprocessing
├── APIC_Reconstruction.py             # GPU-accelerated APIC reconstruction script for basic image reconstruction
├── APIC_Reconstruction_WholeFOV.py    # GPU-accelerated reconstruction script for full-FOV (2560x2560) images, including auto-stitching functionality
├── subfunctionAPIC                    # Directory for APIC subfunctions used in the reconstruction process
├── Sample location segmentation       # Directory for sample location segmentation
    ├── NSCLC.tif                      # Example whole-slide image captured by the sample-locating system
    ├── Sample_Segmentation.py         # Code for sample location segmentation
└── README.md                          # Project documentation (this file)

Table of Contents

• Introduction
• Installation
• Usage
• BiBTeX

Introduction

This repository contains the codes and demonstration data for a Whole Slide Imaging system based on Angular Ptychographic Imaging with a Closed-form Solution (WSI-APIC). WSI-APIC utilizes Segment Anything model for initial high-level sample location segmentation from a whole slide image, thereby bypassing unnecessary scanning of the background regions and enhancing image acquisition efficiency. A GPU-accelerated APIC algorithm analytically reconstructs phase images with effective digital aberration corrections and improved optical resolutions.

Installation

1. Clone this repository to your local machine:

   git clone https://github.com/Magishe/WSI-APIC.git

2. Navigate to the project directory:

   cd WSI-APIC

3. Install the dependencies: To set up your environment and install all the necessary packages, run the following command:

   pip3 install torch==2.1.1+cu121 torchvision==0.16.1+cu121 torch-dct==0.1.6 --index-url https://download.pytorch.org/whl/cu121
   pip3 install numpy scipy matplotlib pillow h5py opencv-python torch-dct

Usage

1. Sample location segmentation

Implement Sample_Segmentation.py to automatically locate and segment samples from the image captured by our sample-locating system (Sample location segmentation/NSCLC.tif).

python Sample_Segmentation.py

2. APIC Reconstruction

Implement APIC_Reconstruction.py to perform GPU-accelerated APIC reconstruction on small ROI patches.

python APIC_Reconstruction.py

Tunable Parameters:

(1). Dataloading

Assume we want to reconstruct the Siemens Star sample which was imaged using a highly aberrated imaging system, which is inside a folder named "Data". Then, we modify the code as

  python APIC_Reconstruction.py --folderName 'Data'

As there is only one file inside the reducedData folder whose name contains "Siemens_Star_g", we can set fileNameKeyword with name "Siemens_Star_g". If there are multiple files, then we could use additionalKeyword

  python APIC_Reconstruction.py --folderName 'Data' --fileNameKeyword 'Siemens_Star_g'

(2). Basic parameters

1. enableROI: When it is set to false, the program uses the entire field-of-view in the reconstruction. It is recommended to set to true as APIC scales badly with respect to the patch size. A good practice is conducting reconstruction using multiple patches and stiching them together to obtain a larger reconstruction coverage.
2. ROILength: This parameter is used only when useROI is true. It specifies the patch sizes used in the reconstruction. It is preferable to set this to be below 256.
3. ROIcenter: Define the center of ROI. Example: ROIcenter = [256,256]; ROIcenter = 'auto'.
4. useAbeCorrection: Whether to enable aberration correction. It is always recommended to set to true. We keep this parameter so that one can see the influence of the aberration if we do not take aberration into consideration.
5. paddingHighRes: To generate a high-resolution image, upsampling is typically requried due to the requirement of Nyquist sampling. paddingHighRes tells the program the upsampling ratio.

Demo Usage:

  python APIC_Reconstruction.py --enableROI --ROILength 256 --ROIcenter auto --useAbeCorrection --paddingHighRes 3

3. APIC Reconstruction for WholeFOV

GPU-accelerated reconstruction script for full-FOV (2560x2560) images, including auto-stitching functionality

python APIC_Reconstruction_WholeFOV.py

Tunable Parameters:

1. patchNumber: The total number of patches into which you intend to divide the full field of view (FOV) along one dimension
2. overlappingSize: The overlap size between different patches (for stitching inside one FOV)

Demo Usage:

  python APIC_Reconstruction_WholeFOV.py --patchNumber 5 --overlappingSize 20

BiBTeX

  @article{Zhao:24,
    author = {Shi Zhao and Haowen Zhou and Siyu (Steven) Lin and Ruizhi Cao and Changhuei Yang},
    journal = {Biomed. Opt. Express},
    keywords = {Image metrics; Imaging systems; Imaging techniques; Laser sources; Phase imaging; Printed circuit boards},
    number = {10},
    pages = {5739--5755},
    publisher = {Optica Publishing Group},
    title = {Efficient, gigapixel-scale, aberration-free whole slide scanner using angular ptychographic imaging with closed-form solution},
    volume = {15},
    month = {Oct},
    year = {2024},
    url = {https://opg.optica.org/boe/abstract.cfm?URI=boe-15-10-5739},
    doi = {10.1364/BOE.538148},
  }