{"id":16271611,"url":"https://github.com/tonysy/deep-feature-flow-segmentation","last_synced_at":"2025-10-04T04:32:12.273Z","repository":{"id":33701222,"uuid":"115441001","full_name":"tonysy/Deep-Feature-Flow-Segmentation","owner":"tonysy","description":"Deep Feature Flow for Video Semantic Segmentation","archived":false,"fork":false,"pushed_at":"2022-06-21T21:23:49.000Z","size":241,"stargazers_count":35,"open_issues_count":11,"forks_count":4,"subscribers_count":5,"default_branch":"master","last_synced_at":"2024-12-16T01:01:56.226Z","etag":null,"topics":["artificial-intelligence","deep-learning","mxnet","sematic-segmentation","video"],"latest_commit_sha":null,"homepage":"","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/tonysy.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}},"created_at":"2017-12-26T17:15:41.000Z","updated_at":"2024-10-09T06:01:46.000Z","dependencies_parsed_at":"2022-08-26T11:50:39.476Z","dependency_job_id":null,"html_url":"https://github.com/tonysy/Deep-Feature-Flow-Segmentation","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/tonysy%2FDeep-Feature-Flow-Segmentation","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/tonysy%2FDeep-Feature-Flow-Segmentation/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/tonysy%2FDeep-Feature-Flow-Segmentation/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/tonysy%2FDeep-Feature-Flow-Segmentation/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/tonysy","download_url":"https://codeload.github.com/tonysy/Deep-Feature-Flow-Segmentation/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":231742829,"owners_count":18419857,"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","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":["artificial-intelligence","deep-learning","mxnet","sematic-segmentation","video"],"created_at":"2024-10-10T18:14:09.248Z","updated_at":"2025-10-04T04:32:11.761Z","avatar_url":"https://github.com/tonysy.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Deep Feature Flow for Video Semantic Segmentation\nBased on Deeplab V2\n\n## 1. Setup environment\n- If you use our dockerfile, you can run the code easily.\n- If you want to set up your own env, please follow these steps:\n    - We only support `python2.7` now\n    - Install tk: `sudo apt-get -y install python-tk`\n    - Install OpenCV 3.4.1\n    - Install needed python packages with `pip install -r requirements.txt`\n        - If you are in China Mainland, you can use these to speedup\n        `pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple`\n- Then `sh init.sh` to build the lib for faster-rcnn\n    Because we use the code from Deformable ConvNets and the dataloader has some dependencies on faster-rcnn, so you need to build the lib first.\n## 2. Prepare Data and Pretrained Model\n### Cityscapes Data\nYou need to download the cityscapes data from the official webpapge and unzip the data\nPut the data into `data/cityscapes`, you can use soft link to set the data path as the following:\n`ln -s Dataset_path ./data/cityscapes`\n\nIf you want to try DFF, you should download cityscapes video data and put it into `data/cityscapes_video`\n\n### Pretrained Model\nDownload pretrained resnet model flow net from [Onedrive](https://onedrive.live.com/?authkey=%21AAXQgYjWim3Iz6w\u0026cid=F371D9563727B96F\u0026id=F371D9563727B96F%21102798\u0026parId=F371D9563727B96F%21102795\u0026action=locate), and put the model into `mode/pretrained_model/`\n```bash\n./model/pretrained_model/resnet_v1_101-0000.params\n./model/pretrained_model/flownet-0000.params\n```\n\n## 3. Train and Test\n### Training Deeplab V2\n`python ./experiments/deeplab/deeplab_train_test.py --cfg ./experiments/deeplab/cfgs/deeplab_resnet_v1_101_cityscapes_segmentation_base.yaml`\n### Training Deeplab V2 Deformable\n`python ./experiments/deeplab/deeplab_train_test.py --cfg ./experiments/deeplab/cfgs/deeplab_resnet_v1_101_cityscapes_segmentation_dcn.yaml`\n### Training DFF Deeplab V2\n`python ./experiments/deeplab_dff/deeplab_dff_train.py --cfg ./experiments/deeplab_dff/cfgs/deeplab_resnet_v1_101_cityscapes_segmentation_video.yaml`\n\n## 4. Performance \nTBD\n## 5. TODO List\n- [x] Add Scripts \n- [ ] Add experiment results\n- [ ] Add support for Deeplab V3+\n- [ ] Add BiSeNet\n## 6. FAQ\n- Program hang if your system opencv is 2.x and your opencv-python is 3.x\n\n## 7. Acknowledgement\n\nThanks for the official deep featuere flow implementation and deeplab implementation from MSRACVER\n- [Deep Feature Flow](https://github.com/msracver/Deep-Feature-Flow)\n- [Deformable ConvNets](https://github.com/msracver/Deformable-ConvNets)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftonysy%2Fdeep-feature-flow-segmentation","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Ftonysy%2Fdeep-feature-flow-segmentation","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftonysy%2Fdeep-feature-flow-segmentation/lists"}