{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/differentiated-attention-guided-network-over","title":"Differentiated Attention Guided Network Over Hierarchical and Aggregated Features for Intelligent UAV Surveillance","arxiv_id":null,"date":"2023-01-17","proceeding":"IEEE Transactions on Industrial Informatics 2023 1","authors":["Houzhang Fang","Zikai Liao","Xuhua Wang","Yi Chang","and Luxin Yan"],"abstract":"Intelligent unmanned aerial vehicle (UAV)\r\nsurveillance based on infrared imaging has wide applications\r\nin the anti-UAV system for protecting urban security\r\nand aerial safety. However, weak target features and\r\ncomplex background distraction pose great challenges for\r\nthe accurate detection of UAVs. To address this issue, we\r\npropose a novel differentiated attention guided network to\r\nadaptively strengthen the discriminative features between\r\nUAV targets and complex background. First, a novel spatialaware\r\nchannel attention (SCA) is introduced into deep layers\r\nvia preserving critical spatial features and leveraging\r\nchannel interdependencies to focus on the large-scale targets.\r\nThe channel-modulated deformable spatial attention\r\nis introduced into shallow layers via refining channel context\r\nand dynamically perceiving the spatial features for focusing\r\non the small-scale targets. A combination of the\r\nabove two attention mechanisms is employed in intermediate\r\nlayers of the network for concentrating on the mediumscale\r\ntargets. Then, we embed a feature aggregator at the\r\ndetection branches to guide the information exchange of\r\nhigh-level feature maps and low-level feature maps with a\r\nbottom-up context modulation, and integrate an SCA at the\r\nend to further boost the distinctive feature representation\r\nfor task-awareness. The above design can adaptively enhance\r\nmultiscale UAV target features and suppress complex\r\nbackground interferences, leading to better detection\r\nperformance, especially for small targets. Extensive experiments\r\non real infrared UAV datasets reveal that the proposed\r\nmethod outperforms the baseline object detectors by a large margin, validating its feasibility in real-world\r\ninfrared UAV detection. The source code can be found at\r\nhttps://github.com/KALEIDOSCOPEIP/DAGNet.","url_abs":"https://ieeexplore.ieee.org/document/10018470","url_pdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10018470","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"differentiated-attention-guided-network-over","repo_url":"https://github.com/kaleidoscopeip/dagnet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null},{"paper_slug":"differentiated-attention-guided-network-over","repo_url":"https://github.com/IVPLaboratory/DAGNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}