Eval tracker

[kangben258]

Scripts are provided for evaling the tracker.

Refer to tools/eval/README.md for usage.

A note on scripts

At the end of the test, two metrics will be obtained, success(0-1)and precision(0-1), success represents the tracker's score of successfully tracking the target, the larger the value of success means the tracker is more robust and better. precision represents the tracker's accuracy score, the larger the value of precision means the tracker is more accurate.

Here are the test results for DaSiamRPN

model	success	precision
DaSiamRPN	0.322	0.409

[zihaomu]

Please clean the meanfulless comment code.

[kangben258]

Scripts are provided for evaling the tracker.
Refer to tools/eval/README.md for usage.
At the end of the test, two metrics will be obtained, success(0-1) and precision(0-1), success represents the tracker's score of successfully tracking the target, the larger the value of success means the tracker is more robust and better. precision represents the tracker's accuracy score, the larger the value of precision means the tracker is more accurate.
The scores of DaSiamRPN are 0.322 for success and 0.409 for precision.

[fengyuentau]

Is this link provided by otb official or your google drive?

[fengyuentau]

At the end of the test, two metrics will be obtained, success(0-1) and precision(0-1), success represents the tracker's score of successfully tracking the target, the larger the value of success means the tracker is more robust and better. precision represents the tracker's accuracy score, the larger the value of precision means the tracker is more accurate.

You can place this along with dataset information in tools/eval/readme.md.

The scores of DaSiamRPN are 0.322 for success and 0.409 for precision.

Please also add this in the model information in models/object_tracking_dasiamrpn/readme.md.

[ryan1288]

@fengyuentau I'm thinking of taking this incomplete evaluation to the finish line. If that sounds good, I'll understand the current change before proposing work to finish the PR.

[fengyuentau]

@fengyuentau I'm thinking of taking this incomplete evaluation to the finish line. If that sounds good, I'll understand the current change before proposing work to finish the PR.

Sure, feel free to take this.