{"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/zero-shot-learning-of-3d-point-cloud-objects","title":"Zero-shot Learning of 3D Point Cloud Objects","arxiv_id":"1902.10272","date":"2019-02-27","proceeding":null,"authors":["Ali Cheraghian","Shafin Rahman","Lars Petersson"],"abstract":"Recent deep learning architectures can recognize instances of 3D point cloud\nobjects of previously seen classes quite well. At the same time, current 3D\ndepth camera technology allows generating/segmenting a large amount of 3D point\ncloud objects from an arbitrary scene, for which there is no previously seen\ntraining data. A challenge for a 3D point cloud recognition system is, then, to\nclassify objects from new, unseen, classes. This issue can be resolved by\nadopting a zero-shot learning (ZSL) approach for 3D data, similar to the 2D\nimage version of the same problem. ZSL attempts to classify unseen objects by\ncomparing semantic information (attribute/word vector) of seen and unseen\nclasses. Here, we adapt several recent 3D point cloud recognition systems to\nthe ZSL setting with some changes to their architectures. To the best of our\nknowledge, this is the first attempt to classify unseen 3D point cloud objects\nin the ZSL setting. A standard protocol (which includes the choice of datasets\nand the seen/unseen split) to evaluate such systems is also proposed. Baseline\nperformances are reported using the new protocol on the investigated models.\nThis investigation throws a new challenge to the 3D point cloud recognition\ncommunity that may instigate numerous future works.","url_abs":"http://arxiv.org/abs/1902.10272v1","url_pdf":"http://arxiv.org/pdf/1902.10272v1.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":"zero-shot-learning-of-3d-point-cloud-objects","repo_url":"https://github.com/ali-chr/ZSL_3D_point_cloud","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"zero-shot-learning","task_name":"Zero-Shot Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1902.10272","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}