LR-FPN: Enhancing Remote Sensing Object Detection with Location Refined Feature Pyramid Network

LR-FPN: Enhancing Remote Sensing Object Detection with Location Refined Feature Pyramid Network

2 Apr 2024 | Hanqian Li1, Ruinan Zhang1, Ye Pan2, Junchi Ren3, Fei Shen4 5*
The paper introduces a novel Location Refined Feature Pyramid Network (LR-FPN) to enhance the extraction of shallow positional information and facilitate fine-grained context interaction in remote sensing object detection. LR-FPN consists of two primary modules: the Shallow Position Information Extraction Module (SPIEM) and the Contextual Interaction Module (CIM). SPIEM extracts positional and saliency information from low-level feature maps, while CIM injects this information into different layers of the original FPN through spatial and channel interaction, enhancing the object area. Extensive experiments on two large-scale remote sensing datasets (DOTA1.0 and HRSC2016) demonstrate that LR-FPN outperforms state-of-the-art object detection approaches, significantly improving performance in remote sensing scenes. The paper also includes ablation studies and visualizations to validate the effectiveness of the proposed modules.The paper introduces a novel Location Refined Feature Pyramid Network (LR-FPN) to enhance the extraction of shallow positional information and facilitate fine-grained context interaction in remote sensing object detection. LR-FPN consists of two primary modules: the Shallow Position Information Extraction Module (SPIEM) and the Contextual Interaction Module (CIM). SPIEM extracts positional and saliency information from low-level feature maps, while CIM injects this information into different layers of the original FPN through spatial and channel interaction, enhancing the object area. Extensive experiments on two large-scale remote sensing datasets (DOTA1.0 and HRSC2016) demonstrate that LR-FPN outperforms state-of-the-art object detection approaches, significantly improving performance in remote sensing scenes. The paper also includes ablation studies and visualizations to validate the effectiveness of the proposed modules.
Reach us at info@study.space
[slides] LR-FPN%3A Enhancing Remote Sensing Object Detection with Location Refined Feature Pyramid Network | StudySpace