countPHICS2 - Colony Counter Interface

A Python-based graphical interface for automated cell colony counting and analysis using FIJI/ImageJ. This tool provides a user-friendly GUI for processing images of colonies with advanced image analysis, statistical visualization, and batch processing capabilities.
Originally developed by Beata Brzozowska and published DOI: 10.1007/s00411-018-00772-z.

Overview

Count and Plot HIstograms of Colony Size 2 (countPHICS2) is designed to automate the tedious process of counting colonies. It combines the power of FIJI/ImageJ's image processing capabilities with Python's data analysis and visualization tools through an intuitive desktop interactive interface.

Key Features

• User-friendly GUI - Clean, modern interface built with PySide6
• Automated colony detection - Advanced image processing with customizable parameters
• Batch processing - Process multiple colony images automatically
• Statistical analysis - Weibull distribution fitting, KS tests, and descriptive statistics
• Data visualization - Automatic generation of histograms, boxplots, and distribution plots
• Comprehensive reporting - Detailed summary files with metadata and parameters

Requirements

Software Dependencies

• Python 3.12+ - With the following packages:
  • PySide6 (Qt6 bindings)
  • pandas
  • numpy
  • matplotlib
  • seaborn
  • scipy
  • natsort

Installation

Manual Download and Setup

1. FIJI/ImageJ Installation

   • The project comes prepackaged with an installation of ImageJ, so no prior installation is necessary.

2. Install Python dependencies (if not using executable)

   pip install PySide6 pandas numpy matplotlib seaborn scipy natsort

3. Clone or download this repository

   git clone <repository-url>
   cd countPHICS2

Project Structure

countPHICS2/
├── countPHICS.bat         # Windows launcher
├── main_window.py         # Main GUI application
├── ImageJ/
│   ├── ImageJ-win64.exe   # Bundled Windows FIJI/ImageJ executable
│   └── macros/
│       └── macro_moj.py   # FIJI/ImageJ macro (Jython)
├── libraries/
│   ├── grapher.py         # Data visualization and analysis
│   └── layout.qss         # Qt stylesheet
├── assets/
│   └── countphics2.ico    # Application icon
└── README.md              # This file

Usage

Starting the Application

Double click batch file in downloaded directory or run terminal command:

python main_window.py

Workflow

1. Select input directory

   • Select directory images 300-1200 dpi with high contrast for best analysis
     • Described in further detail in original countPHICS paper
     • Example image:
       
   • Click "Browse Input" to select a folder containing your .tif, .tiff, .png, .jpeg or .jpg images
   • Images should be RGB.

2. Choose output location (optional)

   • Click "Browse Output" to specify where results should be saved
   • If left blank, results will be saved in a countPHICS_output folder within the input directory

3. Configure analysis settings

   General Settings:

   • Use same ROI for all images - Define region of interest once, apply to all images
   • Split Image - Crop supported composite TIFF scans into separate dish images
   • Generate plots after processing - Automatically create visualization plots

   Advanced Settings :

   • Rolling Ball Radius - Background subtraction parameter
   • Min Colony Size - Minimum area to count as colony
   • Max Colony Size - Maximum area to count as colony
   • Min Circularity - Shape filter (0.0 = any shape, 1.0 = perfect circle)
   • Sigma - Gaussian blur strength before detection
   • ROI Thickness - Line thickness when selecting plate ROI

4. Launch processing

   • Click "▶ LAUNCH FIJI" to start analysis
   • Monitor progress in the console window
   • FIJI will open and process images automatically

5. Review results

   • Results are saved in the user defined output directory with the following structure:

     countPHICS_output/
     ├── Summary_<cell-line>.txt        # Cell-line-specific results table
     ├── plots/                         # Visualization plots
     │   ├── all_colony_counts_boxplot.png
     │   ├── all_colony_sizes_violinplot.png
     │   └── *_area_hist.png
     ├── size_distribution_files/       # Per-image colony sizes
     │   └── *_size_distribution.txt
     └── image_outputs/                 # Annotated images
         └── *_counted.jpg
     

Output Files

Summary_<cell-line>.txt

The summary filename is derived from the Cell line field in the GUI. For example, entering TP53KO produces Summary_TP53KO.txt. The plotting workflow loads this same cell-line-specific file after FIJI finishes.

Tab-separated values file containing:

• Metadata - Macro version, run timestamp, parameters
• Per-image results:
  • Image name
  • Group (assigned in the group-definition dialog)
  • Number of colonies
  • Min/max counted colony size
  • Median colony size
  • Geometric mean colony size

Example:

# countPHICS2 v2.3.0 run: 2026-02-11 11:30:22

# Parameters: 
# AutoThreshold= False
# SameROI=True
# Image ROI=Roi[Oval, x, y, width, height, pos]
# RollingBall=62
# MinColony=150
# MaxColony=10000
# Circularity=0.75
# Sigma=3.0
# MinThresh=0.0
# MaxThresh=240.0

ImageName\tGroup\tColonies\tMinCountedSize\tMaxCountedSize\tMedianSize\tGeomMeanSize
sampleWT_01.tif\t0\t145\t152\t8234\t1024\t987.42
sampleWT_02.tif\t0\t138\t156\t7891\t1012\t953.18
sampleKO_01.tif\t5\t89\t150\t5621\t894\t915.23
sampleKO_02.tif\t5\t94\t153\t5963\t942\t907.42

Size Distribution Files

Individual .txt files for each image containing:

• Colony count header
• Each colony with its respective area per line (in pixels²)

Example:

Number of colonies: 110
Colony Area
328.0
678.0
175.0
160.0
...

Image Outputs

Individual .jpg files for each image containing:

• Thresholded image of colonies.
• Counted colonies surrounded by green outline.

Output Plots

The grapher.py module provides powerful analysis capabilities using matplotlib and seaborn:

• Boxplot summarizing colony counts among replicates in each assigned group.

• Violin plot summarizing colony size distributions in each assigned group.

• Individual histograms for each image summarizing size ditribution:

Available Plot Types

• histogram() - Distribution with Weibull fit and KS test
• boxplot() - Group comparisons with overlaid strip plots
• scatter() - Correlation analysis
• violin() - Distribution shape visualization

Statistical Features

• Weibull distribution fitting - Maximum likelihood estimation for colony size distributions
• Kolmogorov-Smirnov test - Goodness-of-fit assessment with p-value
• Descriptive statistics - Mean, median, quartiles, std deviation
• Metadata annotation - Plots include macro version and runtime information

Image Processing Pipeline

The FIJI macro (macro_moj.py) performs the following steps:

1. Channel selection - Automatically selects RGB channel with highest contrast
2. Gaussian blur - Noise reduction (sigma parameter)
3. Background subtraction - Rolling ball algorithm
4. ROI definition - Region selection on initial image
5. Thresholding - Set thresholding based on initial image
6. Particle analysis - Size and circularity filtering
7. Result compilation - Colony counting and area measurements

Troubleshooting

General

• For general troubleshooting, output images and summary files provide a quick way to diagnose errors.
• If errors persist please refer below for common errors and how to solve them.

FIJI Path Issues

• Problem: Application can't find FIJI
• Solution: Manually set the FIJI path in the code or parameter file
• On Windows, check: C:\Program Files\Fiji.app\ or C:\Users\<username>\Fiji.app\
• On macOS, check: /Applications/Fiji.app/

No Colonies Detected

• Problem: All images show 0 colonies
• Solutions:
  • Decrease minimum colony size
  • Increase sigma value for more blur
  • Check that images are RGB format

Too Many False Positives

• Problem: Background noise being counted
• Solutions:
  • Increase minimum colony size
  • Increase minimum circularity
  • Adjust rolling ball radius
  • Refine ROI to exclude edges

Memory Errors with Large Batches

• Problem: FIJI crashes or freezes
• Solutions:
  • Process images in smaller batches
  • Increase FIJI's memory allocation in preferences
  • Reduce image resolution if appropriate

Technical Details

Parameter Calculations

Advanced settings parameters are automatically calculated based on image dimensions described in the original paper:

• Rolling ball radius: 0.0306 × image_width
• Min colony size: 0.01 × image_width
• Max colony size: image_width
• Sigma: 0.001 × image_width (if units unknown)
  • Or: (1.9e-6) × dpi² + (6.3e-4) × dpi + 1.3 (if units known)

Weibull Distribution

Colony size distributions often follow Weibull distributions. The grapher fits a two-parameter Weibull distribution using maximum likelihood estimation:

• Shape parameter (k): Controls distribution shape
• Scale parameter (λ): Characteristic size
• KS test p-value: Assesses goodness of fit (p < 0.05 suggests good fit)

Citation

If you use countPHICS2 in your research, please cite the original countPHICS paper:

• https://doi.org/10.1007/s00411-018-00772-z

Contributing

Contributions are welcome! Please feel free to submit pull requests or open issues for bugs and feature requests.

Version History

• v2.5.5 (Current GUI)
  • Full Python GUI implementation
  • Advanced plotting with Weibull fitting
  • Enhanced metadata tracking

Support

For questions, issues, or feature requests:

• Open an issue on GitHub
• Contact the authors:
  • Dennis Simpson: dennis@unc.edu
  • Paolo Guerra: pguerra@unc.edu

Acknowledgments

• Original countPHICS software by Beata Brzozowska et al. 2019
  • https://doi.org/10.1007/s00411-018-00772-z