- python3
- pytorch
- numpy
- opencv
- matlab
cd dataset
./download.sh
cd dataset
./extract.sh
cd scripts
python pre_data_otb.py
Then, a file named vot-otb.pkl
will appear in the folder data
. It contains the image file path and ground truth.
SiamRPNBIG.model: https://drive.google.com/file/d/1-vNVZxfbIplXHrqMHiJJYWXYWsOIvGsf/view?usp=sharing
cd DaSiamRPN
python main.py
Then, a file named DaSiamRPN.json
will appear in the folder results
. It records the results.
cd DaSiamRPNAT
python main.py
Then, a file named DaSiamRPNAT.json
will appear in the folder results
.
cd DaSiamRPNAT2
python main.py
Then, a file named DaSiamRPNAT2.json
will appear in the folder results
.
cd scripts
python generate_mat.py -i ../results/DaSiamRPN.json -n DaSiamRPN
python generate_mat.py -i ../results/DaSiamRPNAT.json -n DaSiamRPNAT
python generate_mat.py -i ../results/DaSiamRPNAT2.json -n DaSiamRPNAT2
In the folder otb-toolkit
, run matlab. Then, run run_OPE
in matlab.
cd otb-toolkit
matlab
In matlab:
>> run_OPE
A plot will appear.
- DaSiamRPN: https://github.com/foolwood/DaSiamRPN
- benchmark results: https://github.com/foolwood/benchmark_results
- DAT: https://github.com/shipubupt/NIPS2018
- otb-toolkit: https://github.com/ZhouYzzz/otb-toolkit
- SiamRPN: https://github.com/zkisthebest/Siamese-RPN