{"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/adaptive-o-cnn-a-patch-based-deep","title":"Adaptive O-CNN: A Patch-based Deep Representation of 3D Shapes","arxiv_id":"1809.07917","date":"2018-09-21","proceeding":null,"authors":["Peng-Shuai Wang","Chun-Yu Sun","Yang Liu","Xin Tong"],"abstract":"We present an Adaptive Octree-based Convolutional Neural Network (Adaptive\nO-CNN) for efficient 3D shape encoding and decoding. Different from\nvolumetric-based or octree-based CNN methods that represent a 3D shape with\nvoxels in the same resolution, our method represents a 3D shape adaptively with\noctants at different levels and models the 3D shape within each octant with a\nplanar patch. Based on this adaptive patch-based representation, we propose an\nAdaptive O-CNN encoder and decoder for encoding and decoding 3D shapes. The\nAdaptive O-CNN encoder takes the planar patch normal and displacement as input\nand performs 3D convolutions only at the octants at each level, while the\nAdaptive O-CNN decoder infers the shape occupancy and subdivision status of\noctants at each level and estimates the best plane normal and displacement for\neach leaf octant. As a general framework for 3D shape analysis and generation,\nthe Adaptive O-CNN not only reduces the memory and computational cost, but also\noffers better shape generation capability than the existing 3D-CNN approaches.\nWe validate Adaptive O-CNN in terms of efficiency and effectiveness on\ndifferent shape analysis and generation tasks, including shape classification,\n3D autoencoding, shape prediction from a single image, and shape completion for\nnoisy and incomplete point clouds.","url_abs":"http://arxiv.org/abs/1809.07917v1","url_pdf":"http://arxiv.org/pdf/1809.07917v1.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":"adaptive-o-cnn-a-patch-based-deep","repo_url":"https://github.com/Microsoft/O-CNN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}