CentripetalNet: Pursuing High-quality Keypoint Pairs for Object Detection

TL;DR

CentripetalNet improves keypoint pairing accuracy in object detection by using centripetal shifts and advanced feature adaptation, leading to state-of-the-art performance on MS-COCO.

Contribution

The paper introduces a novel centripetal shift method for keypoint pairing and a cross-star deformable convolution network for enhanced feature extraction.

Findings

Achieves 48.0% AP on MS-COCO test-dev

Outperforms all existing anchor-free detectors

Attains 40.2% MaskAP for instance segmentation

Abstract

Keypoint-based detectors have achieved pretty-well performance. However, incorrect keypoint matching is still widespread and greatly affects the performance of the detector. In this paper, we propose CentripetalNet which uses centripetal shift to pair corner keypoints from the same instance. CentripetalNet predicts the position and the centripetal shift of the corner points and matches corners whose shifted results are aligned. Combining position information, our approach matches corner points more accurately than the conventional embedding approaches do. Corner pooling extracts information inside the bounding boxes onto the border. To make this information more aware at the corners, we design a cross-star deformable convolution network to conduct feature adaption. Furthermore, we explore instance segmentation on anchor-free detectors by equipping our CentripetalNet with a mask prediction module. On MS-COCO test-dev, our CentripetalNet not only outperforms all existing anchor-free detectors with an AP of 48.0% but also achieves comparable performance to the state-of-the-art instance segmentation approaches with a 40.2% MaskAP. Code will be available at https://github.com/KiveeDong/CentripetalNet.