详细分析如下指令的执行过程
# Evaluate performance on VOT
python ../../tools/eval.py --dataset VOT2018 --tracker_prefix C --result_dir ./test/VOT2018
查阅“$SiamMask/tools/eval.py”
eval.py的文档说明中指出该文档来自如下项目:
https://github.com/StrangerZhang/pysot-toolkit.git
查看这个项目的readme。
pysot-toolkit
The purpose of this repo is to provide evaluation API of Current Single Object Tracking Dataset, including
VOT2016
VOT2018
VOT2018-LT
OTB100(OTB2015)
UAV123
NFS
LaSOT
TrackingNet (Evaluation on Server)
GOT-10k (Evaluation on Server)
python …/…/tools/eval.py --dataset VOT2018 --tracker_prefix C --result_dir ./test/VOT2018
loading VOT2018: 100%|████████████████████████████████| 60/60 [00:00<00:00, 2246.11it/s, zebrafish1]
eval ar: 100%|████████████████████████████████████████████████████████| 2/2 [00:01<00:00, 1.89it/s]
eval eao: 100%|███████████████████████████████████████████████████████| 2/2 [00:02<00:00, 1.27s/it]
| Tracker Name | Accuracy | Robustness | Lost Number | EAO |
| Custom_mask_refine_SiamMask_VOT_LD | 0.598 | 0.234 | 50.0 | 0.421 |
| Custom_mask_refine_SiamMask_VOT | 0.610 | 0.281 | 60.0 | 0.380 |