search

hasanirtiza / Pedestron

[Pedestron] Generalizable Pedestrian Detection: The Elephant In The Room. @ CVPR2021
https://openaccess.thecvf.com/content/CVPR2021/papers/Hasan_Generalizable_Pedestrian_Detection_The_Elephant_in_the_Room_CVPR_2021_paper.pdf
Apache License 2.0
687 stars 159 forks source link

How did you compute the Log average miss rate value. #141

Closed ajay1234567899 closed 2 years ago

ajay1234567899 commented 2 years ago

Hello sir, I wanted to know how you computed the log averaged miss rate value. Did you write scripts of your own or did MM detection provide a tool?

hasanirtiza commented 2 years ago

We computed LAMR by using the official scripts provided by the benchmark publishers. For example have a look at CItyPerson. You can have a look at test_city_person.py and test_euroCity.py on how to call them. Additionally, there was also a PR request which modified the original mmdetection evaluation code.

ajay1234567899 commented 2 years ago

Hello sir,

I could generate results for city persons using the scripts you mentioned.

But for the Caltech dataset, I cannot find the scripts that compute the LAMR. It would be great if you could point me to the same.

Thank you