Dynamic Head: Unifying Object Detection Heads with Attentions

The complex nature of combining localization and classification in object detection has resulted in the flourished development of methods. Previous works tried to improve the performance in various object detection heads but failed to present a unified view... In this paper, we present a novel dynamic head framework to unify object detection heads with attentions. By coherently combining multiple self-attention mechanisms between feature levels for scale-awareness, among spatial locations for spatial-awareness, and within output channels for task-awareness, the proposed approach significantly improves the representation ability of object detection heads without any computational overhead. Further experiments demonstrate that the effectiveness and efficiency of the proposed dynamic head on the COCO benchmark. With a standard ResNeXt-101-DCN backbone, we largely improve the performance over popular object detectors and achieve a new state-of-the-art at 54.0 AP. Furthermore, with latest transformer backbone and extra data, we can push current best COCO result to a new record at 60.6 AP. The code will be released at https://github.com/microsoft/DynamicHead. read more

PDF Abstract CVPR 2021 PDF CVPR 2021 Abstract
No code implementations yet. Submit your code now

Datasets


Results from the Paper


 Ranked #1 on Object Detection on COCO test-dev (using extra training data)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Object Detection COCO minival DyHead (Swin-L, multi scale, self-training) box AP 60.3 # 1
AP50 78.2 # 1
AP75 66.3 # 1
APS 46.5 # 1
APM 64.1 # 1
APL 74.2 # 1
Object Detection COCO minival DyHead (Swin-L, multi scale) box AP 58.4 # 5
AP50 76.8 # 3
AP75 64.2 # 2
APS 44.5 # 3
APM 62.2 # 3
APL 73.2 # 3
Object Detection COCO test-dev DyHead (Swin-L, multi scale, self-training) box AP 60.6 # 1
AP50 78.5 # 1
AP75 66.6 # 1
APS 43.9 # 1
APM 64.0 # 1
APL 74.2 # 1
Object Detection COCO test-dev DyHead (Swin-L, multi scale) box AP 58.7 # 5
AP50 77.1 # 2
AP75 64.5 # 2
APS 41.7 # 2
APM 62.0 # 2
APL 72.8 # 2

Methods


No methods listed for this paper. Add relevant methods here