search

bywang2018 / security-dataset

Apache License 2.0
50 stars 9 forks source link

About SDANet in FasterRCNN #14

Open lilitao0517 opened 3 days ago

lilitao0517 commented 3 days ago

Hi author, I applied the SDANet proposed in the paper to the detection task of Faster RCNN and PIDray, but it brought about a decrease in the detection results, my training framework is mmdet2.0+, is there any difference between the training parameters and the default parameters of mmdet, and if so, can you publish the specific parameter settings? thanks!

bywang2018 commented 3 days ago

Thanks very much for your attention to our work. The parameters involved in the experimental paper are all in the publicly available code. I conduct the experiment on the mmdet 2.3.0， and our approach remains effective.

lilitao0517 commented 2 days ago

Thanks very much for your attention to our work. The parameters involved in the experimental paper are all in the publicly available code. I conduct the experiment on the mmdet 2.3.0， and our approach remains effective.

Thanks very much for your quick and accurate answer, the results of retraining after adjusting the parameters & checking the code are close to the original published paper, thank you again for contributing such meaningful work!