[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-127345-en":3,"doc-seo-127345-105":31,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},127345,962085564381,"Clementine","https://ap-avatar.wpscdn.com/davatar_6f874abed73319feea01a86fa6f0fab8",8,"Research & Report","Efficient Machine Learning-Based Semantic Segmentation Algorithm for Consumer-Grade UAV Remote Sensing - Accepted Version","Transformer self-attention’s quadratic growth in sequence length leads to high computational cost and memory use for high-resolution remote sensing imagery, restricting deployment on consumer-grade UAV systems. The proposed EMLSSA introduces hash clustering attention via locality-sensitive hashing to cluster similar tokens into hash buckets and aggregate them with weighted summation. It also adds frequency multi-layer perceptron to fuse frequency and spatial information, strengthening local feature perception. Experiments show EMLSSA-B4 reduces computational cost by 11.7% on FLAME, PWD, EarthVQA, and Potsdam while keeping segmentation quality comparable to SegFormer-B4.","Please cite the Published Version  \nYe, Hong, Cai, Jijing , Deng, Jiangtao , Wang, Xiaodong , Bashir, Ali Kashif , Fang, Kai  and Wang, Wei (2025) Efficient Machine Learning-Based Semantic Segmentation Algorithm for Consumer-Grade UAV Remote Sensing. IEEE Transactions on Consumer Electronics. pp. 1-14. ISSN 0098-3063  \nDOI: [https://doi.org/10.1109/TCE.2025.3596058](https://doi.org/10.1109/TCE.2025.3596058)  \nPublisher: Institute of Electrical and Electronics Engineers (IEEE)  \nVersion: Accepted Version  \nDownloaded from: [https://e-space.mmu.ac.uk/641813/](https://e-space.mmu.ac.uk/641813/)  \nUsage rights:  Creative Commons: Attribution 4 .0  \nAdditional Information: This is an author accepted manuscript of an article published in IEEE Transactions on Consumer Electronics. This version is deposited with a Creative Commons Attribution 4.0 licence [ [https://creativecommons.org/licenses/by/4.0/](https://creativecommons.org/licenses/by/4.0/)]. The version of record can be found on the publisher’s website.  \nEnquiries:  \nIf you have questions about this document, contact [openresearch@mmu.ac.uk](openresearch@mmu.ac.uk. Please)[. Please](openresearch@mmu.ac.uk. Please) include the URL of the record in e-space. If you believe that your, or a third party’s rights have been compromised through this document please see our Take Down policy (available from [https://www.mmu.ac.uk/library/using-the-library/policies-and-guidelines](https://www.mmu.ac.uk/library/using-the-library/policies-and-guidelines))  \nEfficient Machine Learning-Based Semantic Segmentation Algorithm for Consumer-Grade UAV  \nRemote Sensing  \nHong Ye, Jijing Cai, Jiangtao Deng, Xiaodong Wang, Ali Kashif Bashir, Kai Fang, and Wei Wang  \nAbstract—The computational complexity of the Transformer model grows quadratically with input sequence length. This causes a sharp increase in computational cost and memory consumption for high-resolution remote sensing images. Consequently, its application in consumer-grade unmanned aerial vehicle remote sensing is limited. To address this issue, we propose an efficient machine learning-based semantic segmentation algorithm (EMLSSA). First, EMLSSA incorporates the hash clustering attention (HCAttention) mechanism. It employs the locality-sensitive hashing (LSH) algorithm to group similar features into hash buckets, enabling dynamic token clustering. Subsequently, tokens in the same hash bucket are aggregated by weighted summation. This compresses features and reduces the computational complexity of self-attention. Second, EMLSSA incorporates the frequency multi-layer perceptron (FMLP) mechanism. It combines frequency and spatial domain information, enhancing the ability of the Transformer to perceive local features. Experimental results show that EMLSSA-B4 reduces computational cost by 11.7% on FLAME, PWD, EarthVQA, and Potsdam datasets. Furthermore, it maintains comparable segmentation performance to SegFormer-B4.  \nIndex Terms—Machine Learning, Consumer-Grade UAV, Remote Sensing, Locally Sensitive Hashing, Dynamic Clustering, Frequency Multi-Layer Perceptron.  \nI. INTRODUCTION  \nMachine learning-driven semantic segmentation algorithms endow unmanned aerial vehicle (UAV) systems with robust environmental perception capabilities. Through pixel-level image classification, UAV systems can identify and segment various ground features, enabling a detailed understanding of complex scenes [1] . The use of consumer-grade UAVs can significantly improve the efficiency of land resource monitoring, agricultural management, and post-disaster rescue. However,  \nHong Ye is with the College of Electrical and Information Engineering, Quzhou University, Quzhou 324000, China. (Email: [yeh@qzc.edu.cn](yeh@qzc.edu.cn))  \nJijing Cai, Jiangtao Deng, and Kai Fang are with the College of Mathematics and Computer Science, Zhejiang A&F University, Hangzhou 311300, China. (Email: [Jijingcai19@gmail.com](Jijingcai19@gmail.com), [dengjiangtao07@gmail.com](dengjiangt","cbCaijDvFgE4Z85H","https://ap.wps.com/l/cbCaijDvFgE4Z85H","pdf",5345807,2,1,15,"English","en",105,"# Abstract\n## Proposed approach: EMLSSA\n## Hash clustering attention (HCAttention) and LSH token clustering\n## Frequency multi-layer perceptron (FMLP) for local feature perception\n## Experimental evaluation and results\n## Index terms\n# Introduction\n## Motivation for efficient UAV semantic segmentation\n## Limits of CNNs and motivation for Vision Transformer","[{\"question\":\"Why does the Transformer model face difficulties for high-resolution remote sensing on consumer-grade UAVs?\",\"answer\":\"Its self-attention complexity grows quadratically with input sequence length, which sharply increases computational cost and memory consumption for high-resolution images.\"},{\"question\":\"How does EMLSSA reduce the computational complexity of self-attention?\",\"answer\":\"It uses hash clustering attention with locality-sensitive hashing to group similar tokens into hash buckets, then aggregates tokens in each bucket via weighted summation to compress features.\"},{\"question\":\"What role does frequency multi-layer perceptron (FMLP) play in EMLSSA?\",\"answer\":\"FMLP combines frequency and spatial domain information, improving the Transformer’s ability to perceive local features.\"}]","Efficient Machine Learning-Based Semantic Segmentation Algorithm for Consumer-Grade UAV Remote Sensing - Accepted Version | PDF",1785938402,38,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":29},"efficient-machine-learning-based-semantic-segmentation-algorithm-for-consumer-grade-uav-remote-sensing-accepted-version","",{"@graph":37,"@context":86},[38,54,69],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,48,51],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":20},"https://docshare.wps.com/document/","Document",{"item":49,"name":12,"@type":44,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":44,"position":53},"https://docshare.wps.com/document/efficient-machine-learning-based-semantic-segmentation-algorithm-for-consumer-grade-uav-remote-sensing-accepted-version/127345/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":42,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-22","2026-08-05",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"Why does the Transformer model face difficulties for high-resolution remote sensing on consumer-grade UAVs?","Question",{"text":76,"@type":77},"Its self-attention complexity grows quadratically with input sequence length, which sharply increases computational cost and memory consumption for high-resolution images.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does EMLSSA reduce the computational complexity of self-attention?",{"text":81,"@type":77},"It uses hash clustering attention with locality-sensitive hashing to group similar tokens into hash buckets, then aggregates tokens in each bucket via weighted summation to compress features.",{"name":83,"@type":74,"acceptedAnswer":84},"What role does frequency multi-layer perceptron (FMLP) play in EMLSSA?",{"text":85,"@type":77},"FMLP combines frequency and spatial domain information, improving the Transformer’s ability to perceive local features.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,124,129,132,136],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":47,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":47,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":47,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":122,"slug":123},30,"research-report",{"id":125,"doc_module":4,"doc_module_name":47,"category_name":126,"show_sort_weight":127,"slug":128},9,"Religion & Spirituality",20,"religion-spirituality",{"id":127,"doc_module":4,"doc_module_name":47,"category_name":130,"show_sort_weight":127,"slug":131},"World Cup","world-cup",{"id":133,"doc_module":4,"doc_module_name":47,"category_name":134,"show_sort_weight":133,"slug":135},10,"Lifestyle","lifestyle",{"id":137,"doc_module":4,"doc_module_name":47,"category_name":138,"show_sort_weight":107,"slug":139},19,"General","general"]