Remote Sensing Image Fusion With Task-Inspired Multiscale Nonlocal-Attention Network

Liu, Na; Li, Wei; Sun, Xian; Tao, Ran; Chanussot, Jocelyn

doi:10.1109/lgrs.2023.3254049

Cited by 5 publications

(2 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…FRPNet [65] Scale-varying Feature return pyramid structure SBFPN [66] Scale-varying Bidirectional Feature pyramid structure BMFFN [67] Scale-varying Bidirectional multiscale feature fusion network MFPNet [68] Scale-varying Receiver field blocks(RFBs) SSPNet [69] Conflicting information CAM+SSM SMAG [70] Conflicting information Multi-scale supervision module MSAN [71] Scale-varying Multi-scale activation feature fusion block (MAFB) HA-MHGEN [72] Scale-varying Extract explicit and implicit relationships RRNet [73] Scale-varying Parallel multi-scale attention (PMA) A-MLFFMs [74] Scale-varying Adaptively integrate the multi-level outputs MNAN [75] Scale-varying Enhancing multi-scale targets FE-CenterNet [76] Scale-varying FAS+AGS CDD-Net [77] Scale-varying LCFN+HAPN AdaZoom [78] Scale-varying Variable magnification for adaptive multi-scale detection ZoomInNet [79] Scale-varying Adaptive key distillation point(AKDP) UFPMP-Det [80] Scale-varying Unified foreground packing(UFP) SRAF-Net [81] Scale-varying Context-based deformable (CBD) GSDet [82] Scale-varying Converts GSD regression into a probabilistic estimation process GFA-Net [83] Scale-varying Graph Focusing Process (GFP) maps, the attention-guided module enriches the feature representation and enables more comprehensive and discriminative feature learning. The multi-scale attention network [71] is proposed, which incorporates the multi-scale activation feature fusion block to achieve multi-scale attention.…”

Section: Problem Solved Optimization Strategiesmentioning

confidence: 99%

“…The attention-based multi-level feature fusion modules provide a flexible and adaptive framework for effectively exploiting the hierarchical representations of FPN, enabling more accurate and robust target detection across different scales. A multi-scale nonlocal attention-based network [75], which designed to effectively capture and fuse information at multiple scales, allowing for comprehensive analysis of the scene and enhancing the detection performance for objects of various sizes. Tianjun Shi [76] proposes an anchor-free detector that mines multi-scale contextual information using a feature enhancement module consisting of a feature aggregation structure and an attention generation structure, combined with a coordinate attention mechanism to suppress the interference of false alarms in the scene, thus improving the perception of small objects.…”

Section: Problem Solved Optimization Strategiesmentioning

confidence: 99%

See 1 more Smart Citation

A Survey of Small Object Detection Based on Deep Learning in Aerial Images

Wang

Chen

2023

Preprint

View full text Add to dashboard Cite

Small object detection poses a formidable challenge in the field of computer vision, particularly when it comes to analyzing aerial remote sensing images. Despite the rapid development of deep learning and significant progress in detection techniques in natural scenes, the migration of these algorithms to aerial images has not met expectations. This is primarily due to limitations in imaging acquisition conditions, including small target size, viewpoint specificity, background complexity, as well as scale and orientation diversity. Although the increasing application of deep learning-based algorithms to overcome these problems, few studies have summarized the optimization of different deep learning strategies used for small target detection in aerial images. Therefore, this paper aims to explore the application of deep learning methods for small object detection in aerial images. The primary challenges in small object detection in aerial images will be summarized. Next, a meticulous analysis and categorization of the prevailing deep learning optimization strategies employed to surmount the challenges encountered in aerial image detection is undertaken. Following that, we provide a comprehensive presentation of the object detection datasets utilized in aerial remote sensing images, along with the evaluation metrics employed. Additionally, we furnish experimental data pertaining to the currently proposed detection algorithms. Finally, the advantages and disadvantages of various optimization strategies and potential development trends are discussed. Hopefully, it can provide a reference for researchers in this field.

show abstract