How leequin-04/vehicleattributerecognition works
This is a vehicle-attribute recognition project whose principal runtime workflow reads video, detects and tracks vehicles, stabilizes identities, predicts attributes, and writes an annotated video. The README also describes type, color, and make recognition across three datasets; dataset inspection and preparation scripts support those recognition tasks. The runtime coordinator and core calls are evidenced in excerpts. Some dataset-related modules are represented from their names and the README, but their internal behavior and integration with model training were not inspected.
Main entry points: run_detection_tracking.py Video pipeline · identity_manager.py Identity manager · detector_tracker.py Vehicle detector · attribute_provider.py Attribute provider
Components 16
Video recognition
- Video pipeline
src/detection_tracking_var/run_detection_tracking.py - Vehicle detector
src/detection_tracking_var/detector_tracker.py - Identity manager
src/detection_tracking_var/identity_manager.py
Attribute recognition
- Attribute provider
src/detection_tracking_var/attribute_provider.py
Dataset support
- MIO-TCD preparation
src/prepare_miotcd.py - MIO-TCD analysis
src/analyzed_miotcd.py - MIO-TCD samples
src/show_miotcd_samples.py - Make preparation
src/prepare_vehicle_make.py - Make analysis
src/analyze_vehicle_make.py - Make verification
src/verify_vehicle_make.py - Color analysis
src/analyze_vehicle_color.py
Other components
- Video operator
- Annotated video
- MIO-TCD dataset
- Riotu Cars dataset
- Vehicle Color dataset
Connections 13
- Video operator → Video pipeline: starts
README.md - Video pipeline → Vehicle detector: updates tracks
src/detection_tracking_var/run_detection_tracking.py - Video pipeline → Identity manager: updates identities
src/detection_tracking_var/run_detection_tracking.py - Video pipeline → Identity manager: gets display data
src/detection_tracking_var/run_detection_tracking.py - Identity manager → Attribute provider: predicts attributes
src/detection_tracking_var/identity_manager.py - Video pipeline → Annotated video: writes video
src/detection_tracking_var/run_detection_tracking.py - MIO-TCD dataset → MIO-TCD preparation: reads images
src/prepare_miotcd.py - Riotu Cars dataset → Make preparation: reads images
src/prepare_vehicle_make.py - MIO-TCD dataset → MIO-TCD analysis: inspects dataset
src/analyzed_miotcd.py - MIO-TCD dataset → MIO-TCD samples: shows samples
- Riotu Cars dataset → Make analysis: inspects dataset
src/analyze_vehicle_make.py - Riotu Cars dataset → Make verification: verifies dataset
- Vehicle Color dataset → Color analysis: inspects dataset
src/analyze_vehicle_color.py