How harishs-io/lateral_flowimmunoassay_quant works
This repository is an automated LFIA image-analysis and quantification pipeline. The principal workflow processes assay images, measures control/test signals, aggregates replicates, and produces calibration and analytical-limit reports. It also supports optional YOLO fine-tuning from collected strip samples. The main script excerpt is partial, so some connections are shown as unverified rather than inferred as definite calls.
Main entry points: lfia_analyzer.py Pipeline runner · detector.py Model fine-tuning · best.pt YOLO weights · signal_processing.py LSPR extraction
Components 17
Analysis workflow
- Pipeline runner
lfia_analyzer.py - Assay images
- Concentration parsing
src/utils.py - Replicate aggregation
lfia_analyzer.py
Strip vision
- Strip localization
src/detector.py - YOLO weights
weights/best.pt - Model fine-tuning
src/detector.py - Sample deduplication
src/utils.py - Training dataset
Signal quantification
- LSPR extraction
src/signal_processing.py - Band detection
src/signal_processing.py - Peak integration
src/signal_processing.py
Calibration and reporting
- Hook detection
src/calibration.py - Weighted 4PL fit
src/calibration.py - CLSI limits
src/calibration.py - Plots and reports
lfia_analyzer.py
Other components
- Analyst
Connections 18
- Analyst → Pipeline runner: runs
- Assay images → Pipeline runner: provides images
- Pipeline runner → Strip localization: localizes strips
lfia_analyzer.py - YOLO weights → Strip localization: supplies weights
- Pipeline runner → LSPR extraction: extracts signal
lfia_analyzer.py - LSPR extraction → Band detection: detects bands
src/signal_processing.py - Band detection → Peak integration: integrates peaks
src/signal_processing.py - Pipeline runner → Peak integration: calls integration
lfia_analyzer.py - Pipeline runner → Concentration parsing: cleans values
lfia_analyzer.py - Pipeline runner → Replicate aggregation: aggregates replicates
lfia_analyzer.py - Pipeline runner → Hook detection: checks saturation
lfia_analyzer.py - Pipeline runner → Weighted 4PL fit: fits calibration
lfia_analyzer.py - Pipeline runner → CLSI limits: calculates limits
lfia_analyzer.py - Pipeline runner → Plots and reports: writes results
lfia_analyzer.py - Pipeline runner → Model fine-tuning: offers fine-tuning
lfia_analyzer.py - Model fine-tuning → Sample deduplication: deduplicates samples
src/detector.py - Model fine-tuning → Training dataset: writes training data
src/detector.py - Model fine-tuning → YOLO weights: updates weights
src/detector.py