BBRefinement: an universal scheme to improve precision of box object detectors
We present a conceptually simple yet powerful and flexible scheme for refining predictions of bounding boxes. Our approach can be built on top of an arbitrary object detector and produces more precise predictions. The method, called BBRefinement, uses mixture data of image information and the object's class and center. Due to the transformation of the problem into a domain where BBRefinement does not care about multiscale detection, recognition of the object's class, computing confidence, or multiple detections, the training is much more effective. It results in the ability to refine even COCO's ground truth labels into a more precise form. BBRefinement improves the performance of SOTA architectures up to 2mAP points on the COCO dataset in the benchmark. The process of refinement is fast, able to run in real-time on standard hardware. The code is available at https://gitlab.com/irafm-ai/bb-refinement.
PDF Abstract