PBRnet: Pyramidal Bounding Box Refinement to Improve Object Localization Accuracy

TL;DR

PBRnet is a novel boundary refinement architecture that enhances object localization accuracy by combining coarse-to-fine and feature pyramid strategies, significantly improving detection performance on MS-COCO.

Contribution

The paper introduces PBRnet, a boundary refinement network that leverages multi-resolution features to improve bounding box localization in object detection.

Findings

PBRnet improves mAP by approximately 3 points when added to FPN or Libra R-CNN.

Replacing the localization branch in Cascade R-CNN with PBRnet's regressor yields an additional 1.5 mAP improvement.

Overall, PBRnet achieves up to 5 points increase in mAP on MS-COCO.

Abstract

Many recently developed object detectors focused on coarse-to-fine framework which contains several stages that classify and regress proposals from coarse-grain to fine-grain, and obtains more accurate detection gradually. Multi-resolution models such as Feature Pyramid Network(FPN) integrate information of different levels of resolution and effectively improve the performance. Previous researches also have revealed that localization can be further improved by: 1) using fine-grained information which is more translational variant; 2) refining local areas which is more focused on local boundary information. Based on these principles, we designed a novel boundary refinement architecture to improve localization accuracy by combining coarse-to-fine framework with feature pyramid structure, named as Pyramidal Bounding Box Refinement network(PBRnet), which parameterizes gradually focused boundary areas of objects and leverages lower-level feature maps to extract finer local information when refining the predicted bounding boxes. Extensive experiments are performed on the MS-COCO dataset. The PBRnet brings a significant performance gains by roughly 3 point of $mAP$ when added to FPN or Libra R-CNN. Moreover, by treating Cascade R-CNN as a coarse-to-fine detector and replacing its localization branch by the regressor of PBRnet, it leads an extra performance improvement by 1.5 $mAP$, yielding a total performance boosting by as high as 5 point of $mAP$.