当前位置：网站首页>[CVPR 2022] target detection sota:dino: Detr with improved detecting anchor boxes for end to end object detection

[CVPR 2022] target detection sota:dino: Detr with improved detecting anchor boxes for end to end object detection

2022-07-06 23:55:00 【Bit reachable duck】

DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection

Abstract ：
Brief introduction of the paper ：
Concrete realization ：
experimental result ：

This article is the third of the team （ The first two articles are respectively DAB-DETR [ICLR 2022] and DN-DETR [CVPR 2022], Realized DETR Model in COCO For the first time on SOTA

Abstract ：

This paper proposes DINO（DETR with Improved deNoising anchOr boxes） Use ResNet50 As a backbone network, it can be used in COCO Training 12 Wheel can reach 48.3 mAP,36 Wheel can reach 51.0 mAP. And others on the leaderboard

原网站

版权声明
本文为[Bit reachable duck]所创，转载请带上原文链接，感谢
https://yzsam.com/2022/187/202207061632457325.html