JarvisUSTC / RotatedObjectDetectionEvaluation

This repo is evaluation codes for rotated object detection which uses polygon representation (x1,y1,x2,y2...,x4,y4) of bounding box, mainly based on Deformable DETR.
MIT License
7 stars 3 forks source link

RotatedObjectDetectionEvaluation

Summary

This repo is evaluation codes for rotated object detection which uses polygon representation (x1,y1,x2,y2...,x4,y4) for object, mainly based on Deformable DETR.

Why I create this repo

I know in detectron2 RotatedCOCOEvaluator can satisfy almost all cases for rotated object detection. But it has two limitations. First, it only supports angle representation for rotated rectangle, like (cx,cy,w,h,\theta). But in some cases, the groundtruth is not rotated rectangle and we need polygon representation. Secondly, in some times we need to discuss the testing data using data augmentations, such as rotated data augmentation to explore the model generality. But these evaluation methods I can find are for no change testing dataset. So I rewrite the COCO evaluator to support polygon representation and rewrite the evaluation method of DeformableDetr to support data augmentation for testing data.

Usage

Link

https://github.com/fundamentalvision/Deformable-DETR

https://github.com/cocodataset/cocoapi