got-10k / toolkit

Official Python toolkit for generic object tracking benchmark GOT-10k and beyond
http://got-10k.aitestunion.com/
MIT License
555 stars 95 forks source link

Accumulated improvement #24

Closed MARMOTatZJU closed 4 years ago

MARMOTatZJU commented 5 years ago

Several improvements which could be good for utilization.

got10k/datasets/got10k.py

Add option "check_integrity" for the case the user only needs to plot curves.

got10k/datasets/lasot.py

Add three subset "train_test" for Protocol I evaluation purpose.

got10k/datasets/trackingnet.py

Fix TrackingNet dataset bug.

got10k/experiments/init.py

Add TrackingNet & LaSOT experiments.

got10k/experiments/got10k.py

Add option "use_dataset" to skip integrity check in case that the user only need to plot curves. Also add video save functionality

got10k/experiments/lasot.py

Add LaSOT experiment

got10k/experiments/otb.py

Add plot curves

got10k/experiments/trackingnet.py

Add TrackingNet experiment

got10k/utils/metrics.py

Add TrackingNet Normalized Precision score metrics.