search

chenhang98 / BPR

code for `Look Closer to Segment Better: Boundary Patch Refinement for Instance Segmentation`
Apache License 2.0
173 stars 23 forks source link

Getting AP values for each category #37

Closed lavenda-zhou closed 1 year ago

lavenda-zhou commented 1 year ago

Hello, author. How do I get AP values for each category？I have four different types of vehicles in my dataset and want to know the AP value of each. I look forward to your reply, which is very important to me. thank you!

chenhang98 commented 1 year ago

You can refer to https://github.com/open-mmlab/mmdetection/blob/f8bbba2fe777dd882a8da20625be858a98436703/mmdet/datasets/coco.py#L540

lavenda-zhou commented 1 year ago

Thank you for your reply. I have solved this problem. Now I find that the patch size is 6464. Your paper shows that the best input size is 256256. Where can I change this parameter? Thank you for your answering.

chenhang98 commented 1 year ago

The patch size and input size are not necessarily equal in our work. Simply put, we extract several 64x64 patches and resize them to 256x256 (upsample 4x) before inputting them to HRNet. We use this setting in both the code and the paper.