📊 Cell Seeding Efficiency Measurement with Cellpose + SAM
This project analyzes live cell imaging time-series data using Cellpose, SAM, and Microsam to classify cell states and compute seeding efficiency metrics.
---
📁 Folder Structure
- data/raw/: Original input .tif images
- data/processed/: Segmentations, masks, or extracted features
- results/masks/: Per-frame labeled masks for 4 classes
- results/videos/: AVI visualizations (overlays + masks)
- results/metrics.csv: Table of cell count, area, brightness per frame
- logs/: Daily notes and changelogs
---
🧪 Classification Targets
1. Circular cells (initial frames)
2. Fixed cells (midpoint)
3. Dead circular cells (post-fixation)
4. Fragments or abnormal detections
---
⚙ Workflow
1. Preprocess .tif using Cellpose+SAM
2. Segment and classify per frame
3. Track and extract brightness, area, count
4. Output:
- Masked videos
- Combined overlay videos
- CSV with per-frame stats
---
▶ Output
- metrics.csv: frame, class, avg_brightness, avg_area, count
- AVI videos with overlayed masks and raw frames
---
🧠 Dependencies
Requirements.txt Py packages for the Python environment setup.
---
📆 Logs
Daily logs are maintained in logs/daily.md