{"id":32976600,"url":"https://github.com/htung0101/3d_smpl","last_synced_at":"2025-11-16T08:01:53.044Z","repository":{"id":40636986,"uuid":"121676445","full_name":"htung0101/3d_smpl","owner":"htung0101","description":"Code for Paper: Self-supervised Learning of Motion Capture","archived":false,"fork":false,"pushed_at":"2018-02-15T20:15:37.000Z","size":40,"stargazers_count":88,"open_issues_count":5,"forks_count":14,"subscribers_count":9,"default_branch":"master","last_synced_at":"2024-07-21T07:34:31.217Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/htung0101.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2018-02-15T20:02:58.000Z","updated_at":"2024-05-14T06:11:13.000Z","dependencies_parsed_at":"2022-08-30T01:52:09.864Z","dependency_job_id":null,"html_url":"https://github.com/htung0101/3d_smpl","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/htung0101/3d_smpl","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/htung0101%2F3d_smpl","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/htung0101%2F3d_smpl/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/htung0101%2F3d_smpl/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/htung0101%2F3d_smpl/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/htung0101","download_url":"https://codeload.github.com/htung0101/3d_smpl/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/htung0101%2F3d_smpl/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":284678558,"owners_count":27045646,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","status":"online","status_checked_at":"2025-11-16T02:00:05.974Z","response_time":65,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2025-11-13T06:00:31.220Z","updated_at":"2025-11-16T08:01:53.036Z","avatar_url":"https://github.com/htung0101.png","language":"Python","funding_links":[],"categories":["Computer Vision"],"sub_categories":["Geometry"],"readme":"### Self-supervised Learning of Motion Capture ###\n\nThis is code for the paper:\nHsiao-Yu Fish Tung, Hsiao-Wei Tung, Ersin Yumer, Katerina Fragkiadaki, \n[Self-supervised Learning of Motion Capture](https://arxiv.org/abs/1712.01337), NIPS2017 (Spotlight)\n\nCheck the [project page](https://sites.google.com/view/selfsupervisedlearningofmotion/) for more results.\n\n\n### Content ###\n* Environment setup and Dataset\n* Data preprocessing\n* Pretrained model and small tfrecords\n* Training\n* Citation\n* License\n\n### 1. Environment setup and Dataset  ###\n\n* python\nWe use python2.7.13 from Anaconda and Tensorflow 1.1\n\n* SMPL model:\nWe need rest body template from SMPL model. \n\nYou can download it from [here](http://smpl.is.tue.mpg.de/).\n\n* SURREAL Dataset:\nIf you plan to pretrain or test on surreal dataset. \n\nPlease download surreal from [here](https://github.com/gulvarol/surreal)\n\n* H36M Dataset:\nIf you plan to test on real video with some groundtruth (to evaluate). \n\nPlease download H3.6M Dataset from [here](http://vision.imar.ro/human3.6m/description.php)\n\n### 2. Data preprocessing ###\n* Parse Surreal Dataset into binary files\n\nIn order to speed up the read write for tfrecords, we parse surreal dataset into binary files. \nOpen file \n\n    data/preparsed/main_parse_surreal \n\nand change the data path and output path.\n\n* Build up tfrecords\n\nchange the data path to the path you built in the previous step in \n\n    pack_data/pack_data_bin.py\n\nand run it.\nYou can specify how many examples you want to have in each tfrecords by changing value for num_samples.\nIf \"is_test\" is False, we use sequences generated from actor 1, 5, 6, 7, 8 as training samples.\nIf \"is_test\" is True, we use only sequence \"\" from actor 9 as validation.\nYou can change this split by modifying the \"get_file_list\" function in tfrecords_utils.py\n\n### 3. Pretrained model and small tfrecords ###\n\nYou can downdload a pretrained model using supervision from [here](https://drive.google.com/drive/folders/1MB0ATtSfQ7qbvMq49UPYhP2ubd1BfAK-?usp=sharing)\nsurreal_quo0.tfrecords is a small training data and surreal2_100_test_quo1.tfrecords\n\nNote: To make this code pack, I calculate 2d flow directly from 3d groundtruth during testing.\nBut you should replace this with your own predicted flow and keypoints.\n\n### 4. Train model ###\nopen up pretrained.sh, there is one commend for pretraining using supervision,\nand one commend for finetuning with testing data.\nCommend out the line that you need\n\n\n### Citation ###\nIf you use this code, please cite:\n\n@incollection{NIPS2017_7108,\ntitle = {Self-supervised Learning of Motion Capture},\nauthor = {Tung, Hsiao-Yu and Tung, Hsiao-Wei and Yumer, Ersin and Fragkiadaki, Katerina},\nbooktitle = {Advances in Neural Information Processing Systems 30},\neditor = {I. Guyon and U. V. Luxburg and S. Bengio and H. Wallach and R. Fergus and S. Vishwanathan and R. Garnett},\npages = {5236--5246},\nyear = {2017},\npublisher = {Curran Associates, Inc.},\nurl = {http://papers.nips.cc/paper/7108-self-supervised-learning-of-motion-capture.pdf}\n}\n\n\n\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fhtung0101%2F3d_smpl","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fhtung0101%2F3d_smpl","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fhtung0101%2F3d_smpl/lists"}