Neubauer-Counter

A Head-less Python Tool for Automated Cell Counting in Cropped Hemocytometer Frames

Neubauer-Counter is a single-file Python utility that detects and enumerates cells (or other roughly circular particles) in already-cropped images of a Neubauer or Improved Neubauer counting chamber. It is designed for quick batch processing on any machine that has OpenCV and NumPy—no GUI, no manual clicks, no external dependencies beyond standard wheels. The script can also compute cell concentrations (cells · mL⁻¹) when you supply the dilution factor and the frame’s physical volume.

Features

Feature	What it does
Head-less workflow	No cv2.imshow / selectROI; runs on servers, CI, notebooks
Command-line configurable	All blob-detection and I/O options exposed as flags
Overlay generation	Optionally writes JPEG overlays with red dots & indices
CSV summary	Saves a per-image report (cells and, if requested, concentration)
Flexible blob filter	Area range, circularity gate can be tuned or disabled
Live preview (optional)	--preview pops real-time windows when you have a desktop

Installation

1. Clone or download the repo

git clone https://github.com/SidSin0809/neubauer-counter.git cd neubauer-counter

2. Install dependencies

pip install --upgrade pip pip install opencv-python numpy

If you need head-less builds (no HighGUI), replace opencv-python with opencv-python-headless

Quick Start

Count and overlay every *.jpg in the folder

python count_cells_frame.py

Count only PNGs, store outputs in ./results, skip overlay images

python count_cells_frame.py --glob "*.png" --out results --no-overlay

Tighter area filter (50-1200 px²) and no circularity gate

python count_cells_frame.py --min-area 50 --max-area 1200 --no-circularity

Also compute concentration (cells · mL⁻¹)

python count_cells_frame.py --dilution 2.0 --frame-vol 1e-4

CLI Options

Flag	Purpose	Default
--glob "*.jpg"	File pattern to scan	"*.jpg"
--out results	Output directory	results
--min-area 30 / --max-area 1500	Blob area filter (px²)	30–1500
--circularity 0.40 / --no-circularity	Roundness threshold	0.40 (ON)
--no-overlay	Don’t write overlay JPEGs	Off (writes)
--preview	Show live windows	Off
--dilution 2.0	Sample dilution factor (×)	None
--frame-vol 1e-4	Volume of one frame (mL)	None