{"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/scaling-diffusion-models-to-real-world-3d","title":"Scaling Diffusion Models to Real-World 3D LiDAR Scene Completion","arxiv_id":"2403.13470","date":"2024-03-20","proceeding":"CVPR 2024 1","authors":["Lucas Nunes","Rodrigo Marcuzzi","Benedikt Mersch","Jens Behley","Cyrill Stachniss"],"abstract":"Computer vision techniques play a central role in the perception stack of autonomous vehicles. Such methods are employed to perceive the vehicle surroundings given sensor data. 3D LiDAR sensors are commonly used to collect sparse 3D point clouds from the scene. However, compared to human perception, such systems struggle to deduce the unseen parts of the scene given those sparse point clouds. In this matter, the scene completion task aims at predicting the gaps in the LiDAR measurements to achieve a more complete scene representation. Given the promising results of recent diffusion models as generative models for images, we propose extending them to achieve scene completion from a single 3D LiDAR scan. Previous works used diffusion models over range images extracted from LiDAR data, directly applying image-based diffusion methods. Distinctly, we propose to directly operate on the points, reformulating the noising and denoising diffusion process such that it can efficiently work at scene scale. Together with our approach, we propose a regularization loss to stabilize the noise predicted during the denoising process. Our experimental evaluation shows that our method can complete the scene given a single LiDAR scan as input, producing a scene with more details compared to state-of-the-art scene completion methods. We believe that our proposed diffusion process formulation can support further research in diffusion models applied to scene-scale point cloud data.","url_abs":"https://arxiv.org/abs/2403.13470v1","url_pdf":"https://arxiv.org/pdf/2403.13470v1.pdf","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":"scaling-diffusion-models-to-real-world-3d","repo_url":"https://github.com/prbonn/lidiff","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"autonomous-vehicles","task_name":"Autonomous Vehicles"},{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"lidar-scene-completion","task_name":"Lidar Scene Completion"}],"methods":[{"method_slug":"diffusion","method_name":"Diffusion"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/lidar-scene-completion-on-semantickitti","task":"Lidar Scene Completion","dataset":"SemanticKITTI","model":"LiDiff (refined)","rank_in_archive_order":1,"of":4,"metrics":{"Chamfer Distance":"0.376","JSD 3D":"0.573","JSD BEV":"0.416","Voxel IoU 0.1m":"13.40","Voxel IoU 0.2m":"22.99","Voxel IoU 0.5m":"32.43"},"uses_additional_data":false},{"leaderboard":"/sota/lidar-scene-completion-on-semantickitti","task":"Lidar Scene Completion","dataset":"SemanticKITTI","model":"LiDiff","rank_in_archive_order":3,"of":4,"metrics":{"Chamfer Distance":"0.434","JSD 3D":"0.564","JSD BEV":"0.444","Voxel IoU 0.1m":"4.67","Voxel IoU 0.2m":"16.79","Voxel IoU 0.5m":"31.47"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2403.13470","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2403.13470"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/prbonn/lidiff","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran":4,"unverified":2},"by_repo_kind":{"official":{"samples":6,"ran":4,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"273dcb9efc924348","entry":"feats_to_coord","repo":"prbonn/lidiff","repo_kind":"official","path":"lidiff/utils/collations.py","file_url":"https://github.com/prbonn/lidiff/blob/HEAD/lidiff/utils/collations.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"273dcb9efc924348"}},{"code_sha256_prefix":"00586b29e0113f0b","entry":"histogram_point_cloud","repo":"prbonn/lidiff","repo_kind":"official","path":"lidiff/utils/histogram_metrics.py","file_url":"https://github.com/prbonn/lidiff/blob/HEAD/lidiff/utils/histogram_metrics.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"00586b29e0113f0b"}},{"code_sha256_prefix":"8823481d17172a2c","entry":"normalize_pcd","repo":"prbonn/lidiff","repo_kind":"official","path":"lidiff/utils/collations.py","file_url":"https://github.com/prbonn/lidiff/blob/HEAD/lidiff/utils/collations.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8823481d17172a2c"}},{"code_sha256_prefix":"b04f5d8053dc9422","entry":"unormalize_pcd","repo":"prbonn/lidiff","repo_kind":"official","path":"lidiff/utils/collations.py","file_url":"https://github.com/prbonn/lidiff/blob/HEAD/lidiff/utils/collations.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b04f5d8053dc9422"}},{"code_sha256_prefix":"21cd8e2f060ce50c","entry":"load_poses","repo":"prbonn/lidiff","repo_kind":"official","path":"lidiff/map_from_scans.py","file_url":"https://github.com/prbonn/lidiff/blob/HEAD/lidiff/map_from_scans.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"21cd8e2f060ce50c"}},{"code_sha256_prefix":"b39e1616e4146e48","entry":"parse_calibration","repo":"prbonn/lidiff","repo_kind":"official","path":"lidiff/map_from_scans.py","file_url":"https://github.com/prbonn/lidiff/blob/HEAD/lidiff/map_from_scans.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b39e1616e4146e48"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}