代码拉取完成，页面将自动刷新
作者:fengyang95 源地址:https://github.com/fengyang95/pyCFTrackers
Python re-implementation of some correlation filter based tracker, and all of these algorithms are implemented based on the official Matlab code. All the code has been tested on Ubuntu 16.04, Python 3.5. I use pysot-toolkit to eval the performance on OTB and VOT.
git clone https://github.com/wwdguu/pyCFTrackers.git && cd pyCFTrackers
export pyCFTrackers=$PWD
pip install -r requirements.txt
cd lib/eco/features/
python setup.py build_ext --inplace
cd ../../..
cd lib/pysot/utils/
python setup.py build_ext --inplace
cd ../../..
export PYTHONPATH=$PWD:$PYTHONPATH
You can follow the instructions in the following repo to get VOT2016,VOT2018 and OTB100 dataset.
trackdat
Then get the json files according to pysot-toolkit
Then put the data in the dataset dir.
cd examples
python cf_demo.py
cd eval
python get_vot2016_result.py
python get_vot2018_result.py
python ope_otb.py
python eval_VOT2016.py
python eval_VOT2018.py
python eval_OTB.py
| OTB100 Success Plot | OTB100 Precision Plot |
|---|---|
 |
 |
| OTB2013 Success Plot | OTB2013 Precision Plot |
|---|---|
 |
 |
VOT2018
---------------------------------------------------------------
| Tracker Name | Accuracy | Robustness | Lost Number | EAO |
---------------------------------------------------------------
| ECO | 0.485 | 0.403 | 86.0 | 0.224 |
| CSRDCF | 0.492 | 0.501 | 107.0 | 0.210 |
| ECO-HC | 0.500 | 0.473 | 101.0 | 0.207 |
| CSRDCF-LP | 0.503 | 0.553 | 118.0 | 0.199 |
| Staple | 0.524 | 0.665 | 142.0 | 0.179 |
| LDES | 0.528 | 0.684 | 146.0 | 0.175 |
| MCCTH-Staple | 0.535 | 0.684 | 146.0 | 0.172 |
| OPENCV-CSRDCF | 0.486 | 0.651 | 139.0 | 0.170 |
| BACF | 0.511 | 0.674 | 144.0 | 0.169 |
| DAT | 0.477 | 0.777 | 166.0 | 0.158 |
| STRCF | 0.483 | 0.679 | 145.0 | 0.152 |
| CN | 0.439 | 1.100 | 235.0 | 0.112 |
| SAMF | 0.499 | 1.147 | 245.0 | 0.110 |
| DSST | 0.492 | 1.222 | 261.0 | 0.107 |
| DSST-LP | 0.512 | 1.260 | 269.0 | 0.103 |
| DCF | 0.463 | 1.246 | 266.0 | 0.099 |
| KCF | 0.463 | 1.330 | 284.0 | 0.094 |
| CSK | 0.418 | 1.386 | 296.0 | 0.090 |
| MOSSE | 0.378 | 1.967 | 420.0 | 0.063 |
---------------------------------------------------------------
VOT2016
---------------------------------------------------------------
| Tracker Name | Accuracy | Robustness | Lost Number | EAO |
---------------------------------------------------------------
| ECO | 0.564 | 0.256 | 55.0 | 0.336 |
| MCCTH-Staple | 0.574 | 0.359 | 77.0 | 0.303 |
| Staple | 0.560 | 0.387 | 83.0 | 0.299 |
| ECO-HC | 0.532 | 0.350 | 75.0 | 0.293 |
| CSRDCF | 0.542 | 0.359 | 77.0 | 0.273 |
| LDES | 0.577 | 0.419 | 90.0 | 0.272 |
| CSRDCF-LP | 0.548 | 0.354 | 76.0 | 0.272 |
| BACF | 0.521 | 0.405 | 87.0 | 0.252 |
| OPENCV-CSRDCF | 0.521 | 0.438 | 94.0 | 0.239 |
| STRCF | 0.520 | 0.415 | 89.0 | 0.239 |
| DAT | 0.474 | 0.503 | 108.0 | 0.232 |
| SAMF | 0.544 | 0.639 | 137.0 | 0.193 |
| DSST-LP | 0.543 | 0.727 | 156.0 | 0.180 |
| CN | 0.468 | 0.653 | 140.0 | 0.178 |
| DSST | 0.531 | 0.732 | 157.0 | 0.177 |
| DCF | 0.474 | 0.704 | 151.0 | 0.171 |
| KCF | 0.469 | 0.718 | 154.0 | 0.171 |
| CSK | 0.433 | 0.886 | 190.0 | 0.139 |
| MOSSE | 0.388 | 1.244 | 267.0 | 0.096 |
---------------------------------------------------------------
Licensed under an MIT license.
此处可能存在不合适展示的内容，页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容，可点击提交进行申诉，我们将尽快为您处理。
马建仓 AI 助手
