[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-86384-en":3,"doc-seo-86384-105":29,"detail-sidebar-cat-0-en-105":82},{"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":11,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},86384,8796095360427,"Lucas Martin","https://ap-avatar.wpscdn.com/davatar_994ba38a5ba835b3df7d355c54d3ed8d",8,"Research & Report","Can We Trust Unreliable Voxels? Exploring 3D Semantic Occupancy Prediction under Label Noise","3D semantic occupancy prediction is central to robotic perception, but real-world voxel annotations are corrupted by structural artifacts and dynamic trailing effects, raising safety concerns about trusting such supervision. The work introduces OccNL, a dedicated benchmark for 3D occupancy under occupancy-asymmetric noise and real-world dynamic trailing noise. It identifies a severe domain gap where image label-noise methods collapse in sparse 3D voxel spaces. DPROcc addresses this with dual-source partial-label reasoning using temporal memory and representation-level structural affinity to suppress noise propagation, improving semantic performance on SemanticKITTI.","Can we Trust Unreliable Voxels?  \nExploring 3D Semantic Occupancy Prediction under Label Noise  \nWenxin Li 1 , Kunyu Peng2 ,3 ,∗ , Di Wen2 , Junwei Zheng2 , Jiale Wei2 , Mengfei Duan 1 , Yuheng Zhang 1 ,  \nRui Fan4 , and Kailun Yang 1 ,∗  \narXiv :2603 .06279v2 [ cs .CV] 12 Jul 2026  \nFig. 1: Comparison of our proposed DPR-Occ with state-of-the-art label-noise learning methods on our OccNL benchmark. The upper part presents examples of semantic occupancy predictions under noisy supervision, including occupancy-asymmetric and real-world dynamic trailing noise. The lower part shows that existing robust learning strategies struggle to alleviate the adverse effects of voxel-level noise, whereas our proposed noise-robust framework consistently improves semantic performance and achieves state-of-the-art mIoU under both synthetic and real-world noise settings, with particularly pronounced gains in extreme noise scenarios.  \nAbstract—3D semantic occupancy prediction is a cornerstone of robotic perception, yet real-world voxel annotations are inherently corrupted by structural artifacts and dynamic trailing effects. This raises a critical but underexplored question: can autonomous systems safely rely on such unreliable occupancy supervision? To systematically investigate this issue, we establish OccNL, the first benchmark dedicated to 3D occupancy under occupancy-asymmetric and dynamic trailing noise. Our analysis reveals a fundamental domain gap: state-of-the-art 2D label noise learning strategies collapse catastrophically in sparse 3D voxel spaces, exposing a critical vulnerability in existing paradigms. To address this challenge, we propose DPROcc, a principled label-noise-robust framework that constructs reliable supervision through dual-source partial label reasoning. By synergizing temporal model memory with representationlevel structural affinity, DPR-Occ dynamically expands and prunes candidate label sets to preserve true semantics while  \nThis work was supported in part by the National Natural Science Foundation of China (Grant No. 62473139), in part by the Hunan Provincial Research and Development Project (Grant No. 2025QK3019), in part by the State Key Laboratory of Autonomous Intelligent Unmanned Systems (the opening project number ZZKF2025-2-10), and in part by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) -SFB 1574 -471687386.  \n1The authors are with the School of Artificial Intelligence and Robotics and the National Engineering Research Center of Robot Visual Perception and Control Technology, Hunan University, China (email: [kailun.yang@hnu.edu.cn](kailun.yang@hnu.edu.cn)).  \n2The authors are with the Institute for Anthropomatics and Robotics, Karlsruhe Institute of Technology, Germany (email: [kunyu.peng@kit.edu](kunyu.peng@kit.edu)).  \n3The author is also with INSAIT, Sofia University “St. Kliment Ohridski”, Bulgaria.  \n4The author is with the State Key Laboratory of Intelligent Autonomous Systems, Tongji University, Shanghai 201804, China.  \n*Corresponding authors: Kailun Yang and Kunyu Peng.  \nsuppressing noise propagation. Extensive experiments on SemanticKITTI demonstrate that DPR-Occ prevents geometric and semantic collapse under extreme corruption. Notably, even at 90% label noise, our method achieves significant performance gains (up to 2.57% mIoU and 13.91% IoU) over existing label noise learning baselines adapted to the 3D occupancy prediction task. By bridging label noise learning and 3D perception, OccNLand DPR-Occ provide a reliable foundation for safety-critical robotic perception in dynamic environments. The benchmark and source code will be made publicly available at [https:](https:)//[github.com/mylwx/OccNL](github.com/mylwx/OccNL).  \nI. INTRODUCTION  \n3D Semantic Occupancy Prediction [1], also referred to as Semantic Scene Completion (SSC), aims to infer a dense voxel-grid representation of the environment by jointly predicting occupancy and semantic labels. As a foundational ca","cbCaips13XSRlDz8","https://ap.wps.com/l/cbCaips13XSRlDz8","pdf",1238790,4,1,"English","en",105,"# Introduction\n## Problem and motivation\n## OccNL benchmark\n## Why existing methods fail\n## Proposed solution (DPROcc)","[{\"question\":\"How does DPROcc improve robustness to label noise compared with prior label-noise learning strategies?\",\"answer\":\"DPROcc builds reliable supervision via dual-source partial label reasoning, combining temporal model memory with representation-level structural affinity to dynamically expand and prune candidate label sets and suppress noise propagation, yielding improved mIoU/IoU even at extreme noise (e.g., 90%).\"}]",1784211402,20,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":77,"head_meta":79,"extra_data":81,"updated_unix":27},"can-we-trust-unreliable-voxels-exploring-3d-semantic-occupancy-prediction-under-label-noise","",{"@graph":35,"@context":76},[36,52,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":20},"https://docshare.wps.com/document/can-we-trust-unreliable-voxels-exploring-3d-semantic-occupancy-prediction-under-label-noise/86384/",{"url":51,"name":13,"@type":53,"author":54,"headline":13,"publisher":56,"fileFormat":59,"inLanguage":23,"description":14,"dateModified":60,"datePublished":61,"encodingFormat":59,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":55},"Person",{"url":40,"name":57,"@type":58},"DocShare","Organization","application/pdf","2026-07-25","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":20},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70],{"name":71,"@type":72,"acceptedAnswer":73},"How does DPROcc improve robustness to label noise compared with prior label-noise learning strategies?","Question",{"text":74,"@type":75},"DPROcc builds reliable supervision via dual-source partial label reasoning, combining temporal model memory with representation-level structural affinity to dynamically expand and prune candidate label sets and suppress noise propagation, yielding improved mIoU/IoU even at extreme noise (e.g., 90%).","Answer","https://schema.org",{"og:url":51,"og:type":78,"og:title":13,"og:site_name":57,"og:description":14},"article",{"robots":80,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":83},[84,88,92,96,101,106,111,114,118,121,125],{"id":21,"doc_module":4,"doc_module_name":45,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":45,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":45,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":112,"slug":113},30,"research-report",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":28,"slug":117},9,"Religion & Spirituality","religion-spirituality",{"id":28,"doc_module":4,"doc_module_name":45,"category_name":119,"show_sort_weight":28,"slug":120},"World Cup","world-cup",{"id":122,"doc_module":4,"doc_module_name":45,"category_name":123,"show_sort_weight":122,"slug":124},10,"Lifestyle","lifestyle",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":97,"slug":128},19,"General","general"]