How to evaluate detection results in my desktop rather than submit to the test server?

[baiyancheng20]

HI, I divide the training dataset into training and val datasets and use the training dataset to train my detector. How do I test the detection results on my desktop? Do I use 'evaluate_object.cpp' to test my detector performance on the val dataset? I want to find good parameters and then submit my detector's results. Thank you!

[zhaoweicai]

I use evaluate_object.cpp to evaluate the detection performance. But the default code is for evaluation on test set. You need change the code a little bit to make it evaluate on val set.

[jinuhjoo]

Hi @zhaoweicai Could you specify in detail on how you altered and used evaluate_object.cpp That is if I need to run this on my validation set, do I need to split the ground truth txt files in label_2? Could you share your evaluate_object.cpp ?

Thanks ~~

[zhaoweicai]

hi @jinuhjoo Send me an email, I can share the evaluation code to you.

[jk1323]

I have a same question as jinuhjoo,can you share share the evaluation code to me. email [email protected]

[jinuhjoo]

@zhaoweicai I kind'a got it by playing a little with the evaluate_object.cpp, but to get detection box results in txt per image, I had to modify a little in the run_mscnn_detection.m file. Is this how you have approached? My e-mail is [email protected] Thanks :-)

[yqyao]

Could you share the evaluation code to me? I also have the question as jinuhjoo. My e-mail is [email protected],Thanks.

[zhaoweicai]

hi @baiyancheng20 @jinuhjoo @jk1323 @yqyao The evaluation script is uploaded. Please check mscnn/examples/kitti_result/ and README.md for how to use it. Thanks!