Focus-and-Detect: A small object detection framework for aerial images

作者：

Highlights：

• An object detection framework is introduced for small objects in aerial images.

• IBS approach is proposed to suppress false detections on overlapping regions.

• Scale adjustment for bounding boxes is achieved with GMM based clustering technique.

• Proposed framework increases small object detection performance in aerial images.

摘要

•An object detection framework is introduced for small objects in aerial images.•IBS approach is proposed to suppress false detections on overlapping regions.•Scale adjustment for bounding boxes is achieved with GMM based clustering technique.•Proposed framework increases small object detection performance in aerial images.

论文关键词：Object detection,Small object detection,Region search,Aerial images

论文评审过程：Received 25 August 2021, Revised 30 December 2021, Accepted 26 February 2022, Available online 9 March 2022, Version of Record 28 March 2022.

论文官网地址：https://doi.org/10.1016/j.image.2022.116675