ROI Finder

Compute the minimal region of interest (ROI) that contains all line segments listed in one or more CSV files.
Supports both 2D and 3D image volumes, optional voxel scaling, safety margins, and clamping to known image shapes.

---

📦 Overview

compute_min_roi.py reads one or more CSVs containing line endpoints and returns the smallest bounding box that encloses them all.

Each CSV row should describe a line segment with endpoint coordinates, using one of these column patterns:

2D CSV headers	3D CSV headers
x1, y1, x2, y2	x1, y1, z1, x2, y2, z2
x_start, y_start, x_end, y_end	x_start, y_start, z_start, x_end, y_end, z_end

If your CSV headers differ, you can map them using the --cols option.

---

🧮 Example usage

# Basic 2D example
python compute_min_roi.py lines.csv

# 3D example with 3-voxel margin and JSON output
python compute_min_roi.py lines3d.csv --margin 3 --out roi.json

# Multiple CSVs
python compute_min_roi.py a.csv b.csv c.csv

# Clamp to a known image shape (Z Y X for 3D, Y X for 2D)
python compute_min_roi.py lines3d.csv --shape 64 512 512 --int

# Custom headers
python compute_min_roi.py my.csv --cols x1=X_start y1=Y_start x2=X_stop y2=Y_stop

# Include voxel size to also report physical ROI
python compute_min_roi.py lines3d.csv --voxel-size 0.3 0.01 0.01