Spheroid Image Processing Pipeline

This repository contains a series of Python scripts that work after one another to process images of spheroidal tissue models, perform segmentation using a YOLO model, post-process the segmentation masks, and analyze image darkness scores. The scripts are numbered in the order in which they should be executed.

Table of Contents

• Overview
• File Descriptions
  • 00PNGExport.py
  • 01Segmentation_Inference.py
  • 02Masks.py
  • 03DarknessScore.py
• Setup and Installation
• Usage Instructions
• Dependencies
• License

Overview

This pipeline is designed to process high-bit TIFF images of spheroids, convert them to PNG format, perform YOLO-based segmentation inference, generate composite images using segmentation masks, and finally analyze the darkness score (i.e., average pixel intensity) of non-masked areas.

The workflow is as follows:

1. PNG Export (00PNGExport.py):
   Inspects and processes a TIFF image from the Acquisition folder, converting its final layer to an 8-bit grayscale PNG. The result is saved in the output folder.

2. Segmentation Inference (01Segmentation_Inference.py):
   Loads a YOLO segmentation model (Segment.pt) from the working directory and runs inference on PNG images from the output folder. The segmentation outputs (binary masks and object statistics CSV) are saved in the seg_output folder.

3. Mask Post-Processing (02Masks.py):
   Combines the raw PNG image (from the output folder) with the corresponding binary segmentation mask (from seg_output/inferenceSeg/mask_crops/1). The composite images—where background areas are replaced with green—are saved in a new folder RawMasks within the mask folder.

4. Darkness Score Analysis (03DarknessScore.py):
   Processes the composite images from seg_output/inferenceSeg/RawMasks to calculate darkness scores (average pixel intensity) for non-green areas. It also generates intermediate images (green mask, inverted mask, grayscale object image, histogram) and saves the analysis results as a CSV file.

File Descriptions

00PNGExport.py

• Purpose:
  • Inspects a TIFF image and prints its properties (format, mode, size, number of frames, pixel range).
  • Processes the TIFF by converting its final layer from 16-bit to an 8-bit grayscale image.
  • Exports the processed image as a PNG to the output folder.
• Input:
  • A TIFF image (e.g., Spheroid.tiff) located in the Acquisition folder.
• Output:
  • Processed PNG image saved in the output folder.

01Segmentation_Inference.py

• Purpose:
  • Loads the YOLO segmentation model (Segment.pt) from the current working directory.
  • Runs segmentation inference on images located in the output folder.
  • Saves segmentation outputs including binary masks and a CSV (seg_circularity.csv) with area, perimeter, and circularity data.
• Input:
  • PNG images from the output folder.
  • Model file Segment.pt (located in the working directory).
• Output:
  • Segmentation outputs saved in the seg_output folder, including binary mask crops in seg_output/inferenceSeg/mask_crops.

02Masks.py

• Purpose:
  • Combines the raw PNG image from the output folder with its corresponding binary segmentation mask from seg_output/inferenceSeg/mask_crops/1.
  • Creates composite images where pixels corresponding to the mask remain as in the raw image, and the rest are replaced with a green background.
  • Saves the composite images in a new folder named RawMasks within seg_output/inferenceSeg/mask_crops/1.
• Input:
  • Raw PNG images from the output folder.
  • Binary mask images from seg_output/inferenceSeg/mask_crops/1.
• Output:
  • Composite images saved in seg_output/inferenceSeg/mask_crops/1/RawMasks.

03DarknessScore.py

• Purpose:
  • Processes the composite images (from seg_output/inferenceSeg/RawMasks) to compute a darkness score (average grayscale intensity) for the non-green areas.
  • Generates intermediate output images: the green mask, inverted mask, a grayscale image of the object, and a histogram of pixel intensities.
  • Saves the analysis results to a CSV file (average_intensities.csv).
• Input:
  • Composite images (PNG) from seg_output/inferenceSeg/RawMasks.
• Output:
  • Intermediate processing images and a CSV file with darkness scores, all saved in a new subfolder (output) under the processed folder.

Note: You may need to adjust the folder path in 03DarknessScore.py to ensure it points to the correct location relative to your working directory.

Setup and Installation

1. Clone the Repository:

   git clone <repository_url>
   cd <repository_folder>