{"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/discovery-of-latent-3d-keypoints-via-end-to","title":"Discovery of Latent 3D Keypoints via End-to-end Geometric Reasoning","arxiv_id":"1807.03146","date":"2018-07-05","proceeding":"NeurIPS 2018 12","authors":["Supasorn Suwajanakorn","Noah Snavely","Jonathan Tompson","Mohammad Norouzi"],"abstract":"This paper presents KeypointNet, an end-to-end geometric reasoning framework\nto learn an optimal set of category-specific 3D keypoints, along with their\ndetectors. Given a single image, KeypointNet extracts 3D keypoints that are\noptimized for a downstream task. We demonstrate this framework on 3D pose\nestimation by proposing a differentiable objective that seeks the optimal set\nof keypoints for recovering the relative pose between two views of an object.\nOur model discovers geometrically and semantically consistent keypoints across\nviewing angles and instances of an object category. Importantly, we find that\nour end-to-end framework using no ground-truth keypoint annotations outperforms\na fully supervised baseline using the same neural network architecture on the\ntask of pose estimation. The discovered 3D keypoints on the car, chair, and\nplane categories of ShapeNet are visualized at http://keypointnet.github.io/.","url_abs":"http://arxiv.org/abs/1807.03146v2","url_pdf":"http://arxiv.org/pdf/1807.03146v2.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":"discovery-of-latent-3d-keypoints-via-end-to","repo_url":"https://github.com/tensorflow/models/tree/master/research/keypointnet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"3d-pose-estimation","task_name":"3D Pose Estimation"},{"task_slug":"pose-estimation","task_name":"Pose Estimation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1807.03146","atlas_url":"https://app.syntology.ai/?focus=1807.03146","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}