{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:NJEARGBX5FYAP5OFJ2DAT5PZJ3","short_pith_number":"pith:NJEARGBX","schema_version":"1.0","canonical_sha256":"6a48089837e97007f5c54e8609f5f94ecb871edfca0f7dcc0341ac6cf39d23d0","source":{"kind":"arxiv","id":"2106.10689","version":3},"attestation_state":"computed","paper":{"title":"NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"NeuS learns high-fidelity surfaces as neural signed distance functions by using a volume rendering formulation that removes first-order geometric bias.","cross_cats":["cs.GR"],"primary_cat":"cs.CV","authors_text":"Christian Theobalt, Lingjie Liu, Peng Wang, Taku Komura, Wenping Wang, Yuan Liu","submitted_at":"2021-06-20T12:59:42Z","abstract_excerpt":"We present a novel neural surface reconstruction method, called NeuS, for reconstructing objects and scenes with high fidelity from 2D image inputs. Existing neural surface reconstruction approaches, such as DVR and IDR, require foreground mask as supervision, easily get trapped in local minima, and therefore struggle with the reconstruction of objects with severe self-occlusion or thin structures. Meanwhile, recent neural methods for novel view synthesis, such as NeRF and its variants, use volume rendering to produce a neural scene representation with robustness of optimization, even for high"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2106.10689","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2021-06-20T12:59:42Z","cross_cats_sorted":["cs.GR"],"title_canon_sha256":"aa90048bef4588513944561c22fca802a78fa7b8e9555d1cdd593cf70c52319f","abstract_canon_sha256":"c1cb897e46a691bdf57e4718b7332b92e3c6fa59a987392e38c9f122c43892e7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:46.710500Z","signature_b64":"nz6zfg6YO1PScX3ZwOyB+uWSJaIrxcp19MIzZBYzgMLEUAqENiF+FhPagEqmn4pPBALoX7YKfdzr7LtU/z4WAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6a48089837e97007f5c54e8609f5f94ecb871edfca0f7dcc0341ac6cf39d23d0","last_reissued_at":"2026-05-17T23:38:46.710032Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:46.710032Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"NeuS learns high-fidelity surfaces as neural signed distance functions by using a volume rendering formulation that removes first-order geometric bias.","cross_cats":["cs.GR"],"primary_cat":"cs.CV","authors_text":"Christian Theobalt, Lingjie Liu, Peng Wang, Taku Komura, Wenping Wang, Yuan Liu","submitted_at":"2021-06-20T12:59:42Z","abstract_excerpt":"We present a novel neural surface reconstruction method, called NeuS, for reconstructing objects and scenes with high fidelity from 2D image inputs. Existing neural surface reconstruction approaches, such as DVR and IDR, require foreground mask as supervision, easily get trapped in local minima, and therefore struggle with the reconstruction of objects with severe self-occlusion or thin structures. Meanwhile, recent neural methods for novel view synthesis, such as NeRF and its variants, use volume rendering to produce a neural scene representation with robustness of optimization, even for high"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We propose a new formulation that is free of bias in the first order of approximation, thus leading to more accurate surface reconstruction even without the mask supervision.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The new volume rendering formulation eliminates geometric bias without introducing compensating errors or requiring additional constraints beyond the image data.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"NeuS introduces a bias-free volume rendering method for signed distance function representations to reconstruct accurate surfaces from 2D images.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"NeuS learns high-fidelity surfaces as neural signed distance functions by using a volume rendering formulation that removes first-order geometric bias.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1fe1a17452bb6c2c553dbdfccf9ca4d9cdfd06a7d19a7c97728365fd6c410e10"},"source":{"id":"2106.10689","kind":"arxiv","version":3},"verdict":{"id":"73404712-e49a-4f3a-9daa-5c182ba27890","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T20:14:24.628814Z","strongest_claim":"We propose a new formulation that is free of bias in the first order of approximation, thus leading to more accurate surface reconstruction even without the mask supervision.","one_line_summary":"NeuS introduces a bias-free volume rendering method for signed distance function representations to reconstruct accurate surfaces from 2D images.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The new volume rendering formulation eliminates geometric bias without introducing compensating errors or requiring additional constraints beyond the image data.","pith_extraction_headline":"NeuS learns high-fidelity surfaces as neural signed distance functions by using a volume rendering formulation that removes first-order geometric bias."},"references":{"count":54,"sample":[{"doi":"","year":2020,"title":"Sal: Sign agnostic learning of shapes from raw data","work_id":"8a5e5241-6a78-4024-a3c3-1bbe2199cb3e","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2009,"title":"Patchmatch: A ran- domized correspondence algorithm for structural image editing","work_id":"3f45ae92-8b9f-415b-a708-16a29e885730","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2001,"title":"A probabilistic framework for space carving","work_id":"3f8e7ee3-5d08-4820-a3e6-02036b3e96ac","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2019,"title":"Z. Chen and H. Zhang. Learning implicit fields for generative shape modeling. In 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 5932–5941, 2019","work_id":"5891b4db-804b-4e5f-83e8-c361b2dfd60e","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2016,"title":"3d-r2n2: A unified approach for single and multi-view 3d object reconstruction","work_id":"8b3fbcdd-07c8-403a-9032-be0b069d7b19","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":54,"snapshot_sha256":"d5c0f5e7c1f8b2a3313ee8c6518f13654e4819a1bc4a0f47c245a84089fbfec1","internal_anchors":3},"formal_canon":{"evidence_count":2,"snapshot_sha256":"54d0f436a9420dd17967295d80bf8d7d78ca2d101ab7965f59ca7fcea3cd57df"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2106.10689","created_at":"2026-05-17T23:38:46.710110+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.10689v3","created_at":"2026-05-17T23:38:46.710110+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.10689","created_at":"2026-05-17T23:38:46.710110+00:00"},{"alias_kind":"pith_short_12","alias_value":"NJEARGBX5FYA","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_16","alias_value":"NJEARGBX5FYAP5OF","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_8","alias_value":"NJEARGBX","created_at":"2026-05-18T12:33:33.725879+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":30,"internal_anchor_count":30,"sample":[{"citing_arxiv_id":"2605.23888","citing_title":"GenRecon: Bridging Generative Priors for Multi-View 3D Scene Reconstruction","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2205.13524","citing_title":"PREF: Phasorial Embedding Fields for Compact Neural Representations","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2402.12797","citing_title":"A Geometric Algorithm for Blood Vessel Reconstruction from Skeletal Representation","ref_index":34,"is_internal_anchor":true},{"citing_arxiv_id":"2406.04301","citing_title":"Neural Surface Reconstruction from Sparse Views Using Epipolar Geometry","ref_index":42,"is_internal_anchor":true},{"citing_arxiv_id":"2411.18966","citing_title":"SVGS: Enhancing Gaussian Splatting Using Primitives with Spatially Varying Colors","ref_index":29,"is_internal_anchor":true},{"citing_arxiv_id":"2605.16258","citing_title":"IVGT: Implicit Visual Geometry Transformer for Neural Scene Representation","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2602.04516","citing_title":"TACO: Temporal Consensus Optimization for Continual Neural Mapping","ref_index":46,"is_internal_anchor":true},{"citing_arxiv_id":"2605.16582","citing_title":"ArtMesh: Part-Aware Articulated Mesh Fields with Motion-Consistent Dynamics","ref_index":37,"is_internal_anchor":true},{"citing_arxiv_id":"2605.16258","citing_title":"IVGT: Implicit Visual Geometry Transformer for Neural Scene Representation","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2509.04276","citing_title":"PAOLI: Pose-free Articulated Object Learning from Sparse-view Images","ref_index":45,"is_internal_anchor":true},{"citing_arxiv_id":"2509.07435","citing_title":"DreamLifting: A Plug-in Module Lifting MV Diffusion Models for 3D Asset Generation","ref_index":28,"is_internal_anchor":true},{"citing_arxiv_id":"2510.09880","citing_title":"Geometry-Aware Scene Configurations for Novel View Synthesis","ref_index":42,"is_internal_anchor":true},{"citing_arxiv_id":"2511.19172","citing_title":"MetroGS: Efficient and Stable Reconstruction of Geometrically Accurate High-Fidelity Large-Scale Scenes","ref_index":42,"is_internal_anchor":true},{"citing_arxiv_id":"2410.03825","citing_title":"MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motion","ref_index":153,"is_internal_anchor":true},{"citing_arxiv_id":"2605.14594","citing_title":"TOPOS: High-Fidelity and Efficient Industry-Grade 3D Head Generation","ref_index":48,"is_internal_anchor":true},{"citing_arxiv_id":"2604.01204","citing_title":"Neural Harmonic Textures for High-Quality Primitive Based Neural Reconstruction","ref_index":60,"is_internal_anchor":true},{"citing_arxiv_id":"2604.02736","citing_title":"THOM: Generating Physically Plausible Hand-Object Meshes From Text","ref_index":64,"is_internal_anchor":true},{"citing_arxiv_id":"2605.03463","citing_title":"First Shape, Then Meaning: Efficient Geometry and Semantics Learning for Indoor Reconstruction","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2605.09644","citing_title":"Attention Itself Could Retrieve.RetrieveVGGT: Training-Free Long Context Streaming 3D Reconstruction via Query-Key Similarity Retrieval","ref_index":41,"is_internal_anchor":true},{"citing_arxiv_id":"2605.09299","citing_title":"LagrangianSplats: Divergence-Free Transport of Gaussian Primitives for Fluid Reconstruction","ref_index":95,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06408","citing_title":"Scalable GPU Construction of 3D Voronoi and Power Diagrams","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2604.22129","citing_title":"PAGaS: Pixel-Aligned 1DoF Gaussian Splatting for Depth Refinement","ref_index":42,"is_internal_anchor":true},{"citing_arxiv_id":"2605.04524","citing_title":"High-Fidelity Single-Image Head Modeling with Industry-Grade Topology","ref_index":150,"is_internal_anchor":true},{"citing_arxiv_id":"2604.19568","citing_title":"SpUDD: Superpower Contouring of Unsigned Distance Data","ref_index":35,"is_internal_anchor":true},{"citing_arxiv_id":"2604.11983","citing_title":"A Geometric Algebra-informed NeRF Framework for Generalizable Wireless Channel Prediction","ref_index":18,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NJEARGBX5FYAP5OFJ2DAT5PZJ3","json":"https://pith.science/pith/NJEARGBX5FYAP5OFJ2DAT5PZJ3.json","graph_json":"https://pith.science/api/pith-number/NJEARGBX5FYAP5OFJ2DAT5PZJ3/graph.json","events_json":"https://pith.science/api/pith-number/NJEARGBX5FYAP5OFJ2DAT5PZJ3/events.json","paper":"https://pith.science/paper/NJEARGBX"},"agent_actions":{"view_html":"https://pith.science/pith/NJEARGBX5FYAP5OFJ2DAT5PZJ3","download_json":"https://pith.science/pith/NJEARGBX5FYAP5OFJ2DAT5PZJ3.json","view_paper":"https://pith.science/paper/NJEARGBX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.10689&json=true","fetch_graph":"https://pith.science/api/pith-number/NJEARGBX5FYAP5OFJ2DAT5PZJ3/graph.json","fetch_events":"https://pith.science/api/pith-number/NJEARGBX5FYAP5OFJ2DAT5PZJ3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NJEARGBX5FYAP5OFJ2DAT5PZJ3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NJEARGBX5FYAP5OFJ2DAT5PZJ3/action/storage_attestation","attest_author":"https://pith.science/pith/NJEARGBX5FYAP5OFJ2DAT5PZJ3/action/author_attestation","sign_citation":"https://pith.science/pith/NJEARGBX5FYAP5OFJ2DAT5PZJ3/action/citation_signature","submit_replication":"https://pith.science/pith/NJEARGBX5FYAP5OFJ2DAT5PZJ3/action/replication_record"}},"created_at":"2026-05-17T23:38:46.710110+00:00","updated_at":"2026-05-17T23:38:46.710110+00:00"}