Exploiting foreground and background separation for prohibited item detection in overlapping X-Ray images

作者：

Highlights：

• A foreground and background separation framework is proposed for prohibited item detection in heavily overlapping X-ray image.

• Method can separate the pure prohibited items from the X-ray images and eliminate the interference of irrelevant information to the detection.

• Its performance is demonstrated on a synthetic dataset and two public datasets. Experiment results verify the effectiveness of our method.

摘要

•A foreground and background separation framework is proposed for prohibited item detection in heavily overlapping X-ray image.•Method can separate the pure prohibited items from the X-ray images and eliminate the interference of irrelevant information to the detection.•Its performance is demonstrated on a synthetic dataset and two public datasets. Experiment results verify the effectiveness of our method.

论文关键词：X-ray imagery,Object detection,Foreground and background separation (FBS),Recursive training

论文评审过程：Received 3 January 2021, Revised 9 August 2021, Accepted 18 August 2021, Available online 20 August 2021, Version of Record 31 August 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108261