{"id":19514217,"url":"https://github.com/weihaox/st-clstm","last_synced_at":"2026-02-10T22:02:32.159Z","repository":{"id":109200829,"uuid":"234305700","full_name":"weihaox/ST-CLSTM","owner":"weihaox","description":"This is the UNOFFICIAL implementation of the ICCV 2019 paper 'Exploiting Temporal Consistency for Real-Time Video Depth Estimation'.","archived":false,"fork":false,"pushed_at":"2021-05-17T14:05:30.000Z","size":274,"stargazers_count":27,"open_issues_count":1,"forks_count":6,"subscribers_count":4,"default_branch":"master","last_synced_at":"2025-07-31T18:34:53.896Z","etag":null,"topics":["3d-convolutional-network","clstm","depth-estimation","depth-prediction","temporal-models"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/weihaox.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2020-01-16T11:30:39.000Z","updated_at":"2024-10-17T13:14:19.000Z","dependencies_parsed_at":null,"dependency_job_id":"e1a989b7-d473-4ca7-8279-cf46d982d3ed","html_url":"https://github.com/weihaox/ST-CLSTM","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/weihaox/ST-CLSTM","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/weihaox%2FST-CLSTM","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/weihaox%2FST-CLSTM/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/weihaox%2FST-CLSTM/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/weihaox%2FST-CLSTM/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/weihaox","download_url":"https://codeload.github.com/weihaox/ST-CLSTM/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/weihaox%2FST-CLSTM/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":29319257,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-02-10T20:44:44.282Z","status":"ssl_error","status_checked_at":"2026-02-10T20:44:43.393Z","response_time":65,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.6:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"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":["3d-convolutional-network","clstm","depth-estimation","depth-prediction","temporal-models"],"created_at":"2024-11-10T23:35:32.539Z","updated_at":"2026-02-10T22:02:32.109Z","avatar_url":"https://github.com/weihaox.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)\n[![Python 3.6](https://img.shields.io/badge/python-3.6-green.svg)](https://www.python.org/)\n[![PyTorch 1.0](https://img.shields.io/badge/pytorch-1.0-green.svg)](https://pytorch.org/)\n\n# Exploiting Temporal Consistency for Real-Time Video Depth Estimation\nThis is the UNOFFICIAL implementation of the paper [***Exploiting Temporal Consistency for Real-Time Video Depth Estimation***](https://arxiv.org/abs/1908.03706), ***ICCV 2019, Haokui Zhang, Chunhua Shen, Ying Li, Yuanzhouhan Cao, Yu Liu, Youliang Yan.***\n\nYou can find official implementation (WITHOUT TRAINING SCRIPTS) [here](https://github.com/hkzhang91/ST-CLSTM).\n\n## Framework\n![](./assert/framework.png)\n\n## Dependencies\n- [Python3.6](https://www.python.org/downloads/)\n- [PyTorch 1.0+](https://pytorch.org/)\n- [NYU Depth v2](https://cs.nyu.edu/~silberman/datasets/nyu_depth_v2.html)\n\n## Pre-processed Data\nWe didn't preprocess data as in the official implementation. Instead, we use the dataset shared by [Junjie Hu](https://github.com/JunjH/Revisiting_Single_Depth_Estimation), which is also used by [SARPN](https://github.com/Xt-Chen/SARPN/blob/master/README.md).\nYou can download the pre-processed data from [here](https://drive.google.com/file/d/1WoOZOBpOWfmwe7bknWS5PMUCLBPFKTOw/view?usp=sharing). \n\nWhen you have downloaded the dataset, run the following command to creat training list.\n```bash\npython create_list_nyuv2.py\n```\n\nYou can also follow the procedure of [ST-CLSTM](https://github.com/hkzhang91/ST-CLSTM) to preprocess the data. It is based on the oficial Matlab [Toolbox](https://cs.nyu.edu/~silberman/datasets/nyu_depth_v2.html). If Matlab is unavailable for you, there is also a [Python Port Toolbox](https://github.com/GabrielMajeri/nyuv2-python-toolbox) for processing the raw dataset by [GabrielMajeri](https://github.com/GabrielMajeri), which contains code for Higher-level interface to the labeled subset, Raw dataset extraction and preprocessing and Performing data augmentation.\n\nThe final folder structure is shown below.\n```\ndata_root  \n|- raw_nyu_v2_250k  \n| |- train  \n| | |- basement_0001a  \n| | | |- rgb  \n| | | | |- rgb_00000.jpg  \n| | | | |_ ...  \n| | | |- depth  \n| | | | |- depth_00000.png  \n| | | | |_ ...  \n| | |_ ...  \n| |- test_fps_30_fl5_end  \n| | |- 0000  \n| | | |- rgb  \n| | | | |- rgb_00000.jpg  \n| | | | |- rgb_00001.jpg  \n| | | | |- ...  \n| | | | |- rgb_00004.jpg  \n| | | |- depth  \n| | | | |- depth_00000.png  \n| | | | |- depth_00001.png  \n| | | | |- ...  \n| | | | |- depth_00004.png  \n| | |- ...  \n| |- test_fps_30_fl4_end  \n| |- test_fps_30_fl3_end\u003e\n```\n## Train\nAs an example, use the following command to train on NYUDV2.\u003cbr\u003e\n\n```bash\nCUDA_VISIBLE_DEVICES=\"0,1,2,3\" python train.py --epochs 20 --batch_size 128 \\\n                        --resume --do_summary --backbone resnet18  --refinenet R_CLSTM_5 \\\n                        --trainlist_path ./data_list/raw_nyu_v2_250k/raw_nyu_v2_250k_fps30_fl5_op0_end_train.json \\\n                        --root_path ./data/ --checkpoint_dir ./checkpoint/ --logdir ./log/\n```\n## Evaluation\nUse the following command to evaluate the trained model on ST-CLSTM [test data](https://github.com/hkzhang91/ST-CLSTM).\u003cbr\u003e\n\n```bash\nCUDA_VISIBLE_DEVICES=\"0\" python evaluate.py --batch_size 1  --backbone resnet18  --refinenet R_CLSTM_5 --loadckpt ./checkpoint/ \\\n                        --testlist_path ./data_list/raw_nyu_v2_250k/raw_nyu_v2_250k_fps30_fl5_op0_end_test.json \\\n                        --root_path ./data/st-clstm/\n```\n## Pretrained Model\nYou can download the pretrained model: [NYUDV2](https://github.com/hkzhang91/ST-CLSTM/tree/master/CLSTM_Depth_Estimation-master/prediction/trained_models).\n\n## Citation\n\n```bibtex\n@inproceedings{zhang2019temporal,\n             title = {Exploiting Temporal Consistency for Real-Time Video Depth Estimation},\n             author = {Haokui Zhang and Chunhua Shen and Ying Li and Yuanzhouhan Cao and Yu Liu and Youliang Yan},\n             conference={International Conference on Computer Vision},\n             year = {2019}   \n} \n```","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fweihaox%2Fst-clstm","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fweihaox%2Fst-clstm","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fweihaox%2Fst-clstm/lists"}