Abstract

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. The code will be released at https://github.com/microsoft/DynamicHead.

Keywords

Computer scienceObject detectionBenchmark (surveying)Object (grammar)Overhead (engineering)Task (project management)Artificial intelligenceRepresentation (politics)Feature (linguistics)Code (set theory)DetectorComputer visionFeature extractionPattern recognition (psychology)Programming languageEngineering

Affiliated Institutions

Related Publications

Publication Info

Year
2021
Type
article
Pages
7369-7378
Citations
809
Access
Closed

External Links

View on DOI.org

Social Impact

Altmetric
PlumX Metrics

Social media, news, blog, policy document mentions

Citation Metrics

809
OpenAlex

Cite This

Xiyang Dai, Yinpeng Chen, Bin Xiao et al. (2021). Dynamic Head: Unifying Object Detection Heads with Attentions. , 7369-7378. https://doi.org/10.1109/cvpr46437.2021.00729

Identifiers

DOI
10.1109/cvpr46437.2021.00729