Balanced single-shot object detection using cross-context attention-guided network

作者：

Highlights：

• A light but effective Cross-context Attention-guided Network is proposed to maximize the balance between accuracy and speed for real-world object detection applications.

• Cross-context Attention Mechanism (CCAM) is established to take the multi-region context information into consideration simultaneously, such as cross-region, adjacent-region, channel region, and spatial region.

• Three attention mechanisms are introduced to guide the network optimization for better learning region focusing.

摘要

•A light but effective Cross-context Attention-guided Network is proposed to maximize the balance between accuracy and speed for real-world object detection applications.•Cross-context Attention Mechanism (CCAM) is established to take the multi-region context information into consideration simultaneously, such as cross-region, adjacent-region, channel region, and spatial region.•Three attention mechanisms are introduced to guide the network optimization for better learning region focusing.

论文关键词：Cross-context attention-guided network,Cross-context attention mechanism,Receptive field attention mechanism,Semantic fusion attention mechanism,Accuracy and speed balance

论文评审过程：Received 7 June 2020, Revised 14 July 2021, Accepted 13 August 2021, Available online 24 August 2021, Version of Record 29 August 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108258