Enhancing Point Cloud Completion with Fine-Grained Geometric Perception

Enhancing Point Cloud Completion with Fine-Grained Geometric Perception | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Enhancing Point Cloud Completion with Fine-Grained Geometric Perception Limin Zhang, Lu Shi, Linna Zhang, Yi Jin, Yidong Li, Yigang Cen, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8474011/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Point clouds are crucial for representing 3D objects, yet they often suffer from incompleteness due to various factors. Traditional methods for point cloud completion focus on global shape integrity but neglect fine-grained geometric features. This study introduces an Enhancing Point Cloud Completion with Fine-Grained Geometric Perception (FGGP-PCC) model that operates across multiple res-olutions. By integrating a Joint Local Multi-Layer Perceptron (JL-MLP) with an attention mechanism, FGGP-PCC effectively captures both local detailsand long-range semantic dependencies. Our approach significantly enhances themodel’s ability to reason about complex geometric complementarity. Experimental results on benchmark datasets such as ShapeNet-55, PCN, and Completion3D demonstrate state-of-the-art performance, with FGGP-PCC achieving an average reduction of 3.45% in Chamfer Distance across different completion levels. The source code is available at https://github.com/ChinaZlm2022/Loss-Edge-Merging-with-Attention. Point cloud completion multi-resolution attention encoder edge-aware feature learning hierarchical attention decoder direction-aware Chamfer Distance Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8474011","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":578315530,"identity":"c15df203-6575-434c-8576-4fc1c7a49237","order_by":0,"name":"Limin Zhang","email":"","orcid":"","institution":"Bejing Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Limin","middleName":"","lastName":"Zhang","suffix":""},{"id":578315531,"identity":"54353f34-4531-4617-9e19-ca70ba93d3d2","order_by":1,"name":"Lu Shi","email":"","orcid":"","institution":"Bejing Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Lu","middleName":"","lastName":"Shi","suffix":""},{"id":578315532,"identity":"b2ed8527-b8b4-4d67-bc19-76d7971e2b4a","order_by":2,"name":"Linna Zhang","email":"","orcid":"","institution":"Guizhou University","correspondingAuthor":false,"prefix":"","firstName":"Linna","middleName":"","lastName":"Zhang","suffix":""},{"id":578315533,"identity":"e8b1f032-3557-4d82-8250-1951ea83e98d","order_by":3,"name":"Yi Jin","email":"","orcid":"","institution":"Beijing Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Yi","middleName":"","lastName":"Jin","suffix":""},{"id":578315534,"identity":"358fe84f-48e2-4fc2-828c-b4049859db98","order_by":4,"name":"Yidong Li","email":"","orcid":"","institution":"Bejing Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Yidong","middleName":"","lastName":"Li","suffix":""},{"id":578315535,"identity":"b8664f1c-d79e-4534-8666-4ea53a61af69","order_by":5,"name":"Yigang Cen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAw0lEQVRIiWNgGAWjYNCCigPMYJqHeC1nDjDzkKaFse0AA/FaDM4vfvi4cN4ddnuJBMYHb9sY5M0JarnxzNh45rZnzDwSCcyGc9sYDHc2ENRyhk2ad9thkBYgo40hweAAYS3sv3nngLUAGURpOd/DxszbALGFmSgtkjfYjKV5jgH9cuZhs+SccxKGGwhp4Tt/+OFnnpo7yeztyQc/vCmzkSdoi8KNBDCdDIydBiAtQUA9EMj3Qwy1I6x0FIyCUTAKRiwAAHvGPuVKWnntAAAAAElFTkSuQmCC","orcid":"","institution":"Bejing Jiaotong University","correspondingAuthor":true,"prefix":"","firstName":"Yigang","middleName":"","lastName":"Cen","suffix":""},{"id":578315536,"identity":"15cf648c-2c0a-4d23-97bf-571c63ad5d2b","order_by":6,"name":"Jian Zhang","email":"","orcid":"","institution":"Beijing Jiaotong University","correspondingAuthor":false,"prefix":"","firstName":"Jian","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2025-12-29 14:39:03","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8474011/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8474011/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":100939099,"identity":"6c431fff-b129-4cd7-99d0-16665458b755","added_by":"auto","created_at":"2026-01-23 04:03:15","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7312,"visible":true,"origin":"","legend":"","description":"","filename":"d82fcd6880b64446b8f946d5b4029ac1.json","url":"https://assets-eu.researchsquare.com/files/rs-8474011/v1/4740e139c0c3b5a4f46f153f.json"},{"id":106959107,"identity":"13f301d8-9bf6-4ef0-894f-ca2160061d08","added_by":"auto","created_at":"2026-04-15 08:46:23","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":849067,"visible":true,"origin":"","legend":"","description":"","filename":"FGGPPCC.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8474011/v1_covered_c7c3eed9-fa00-41da-81ae-bc17acc9d93a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Enhancing Point Cloud Completion with Fine-Grained Geometric Perception","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Point cloud completion, multi-resolution attention encoder, edge-aware feature learning, hierarchical attention decoder, direction-aware Chamfer Distance","lastPublishedDoi":"10.21203/rs.3.rs-8474011/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8474011/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePoint clouds are crucial for representing 3D objects,\u0026nbsp; yet they often suffer from incompleteness due to various factors. Traditional methods for point cloud completion focus on global shape integrity but neglect fine-grained geometric features. This study introduces an Enhancing Point Cloud Completion with Fine-Grained Geometric Perception\u0026nbsp; (FGGP-PCC)\u0026nbsp; model that operates across multiple res-olutions. By integrating a Joint Local Multi-Layer Perceptron (JL-MLP) with an attention mechanism,\u0026nbsp; FGGP-PCC effectively captures both local detailsand long-range semantic dependencies. Our approach significantly enhances themodel’s ability to reason about complex geometric complementarity.\u0026nbsp; Experimental results on benchmark datasets such as ShapeNet-55, PCN,\u0026nbsp; and Completion3D demonstrate state-of-the-art performance, with FGGP-PCC achieving an average reduction of 3.45% in Chamfer Distance across different completion levels.\u003c/p\u003e\n\u003cp\u003eThe source code is available at https://github.com/ChinaZlm2022/Loss-Edge-Merging-with-Attention.\u003c/p\u003e","manuscriptTitle":"Enhancing Point Cloud Completion with Fine-Grained Geometric Perception","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-23 04:03:10","doi":"10.21203/rs.3.rs-8474011/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0c1a8507-77f6-44b1-b573-d340d102182b","owner":[],"postedDate":"January 23rd, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-12T05:39:28+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-23 04:03:10","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8474011","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8474011","identity":"rs-8474011","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}