{"id":13543141,"url":"https://github.com/bgshih/aster","last_synced_at":"2025-04-13T07:49:56.434Z","repository":{"id":33052568,"uuid":"112913274","full_name":"bgshih/aster","owner":"bgshih","description":"Recognizing cropped text in natural images.","archived":false,"fork":false,"pushed_at":"2023-03-25T00:44:11.000Z","size":367,"stargazers_count":732,"open_issues_count":82,"forks_count":196,"subscribers_count":20,"default_branch":"master","last_synced_at":"2025-04-13T07:49:50.435Z","etag":null,"topics":["computer-vision","ocr","recognition","scene-text"],"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/bgshih.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}},"created_at":"2017-12-03T08:50:02.000Z","updated_at":"2025-04-02T13:12:33.000Z","dependencies_parsed_at":"2023-01-14T23:12:25.841Z","dependency_job_id":"9913470e-51e8-411a-857d-c2d80c2b7a56","html_url":"https://github.com/bgshih/aster","commit_stats":null,"previous_names":[],"tags_count":3,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bgshih%2Faster","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bgshih%2Faster/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bgshih%2Faster/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bgshih%2Faster/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/bgshih","download_url":"https://codeload.github.com/bgshih/aster/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248681494,"owners_count":21144700,"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":["computer-vision","ocr","recognition","scene-text"],"created_at":"2024-08-01T11:00:24.007Z","updated_at":"2025-04-13T07:49:56.409Z","avatar_url":"https://github.com/bgshih.png","language":"Python","funding_links":[],"categories":["Text detection and localization"],"sub_categories":["Form Segmentation"],"readme":"# ASTER: Attentional Scene Text Recognizer with Flexible Rectification\n\nASTER is an accurate scene text recognizer with flexible rectification mechanism. The research paper can be found [here](https://ieeexplore.ieee.org/abstract/document/8395027/).\n\n![ASTER Overview](overview.png)\n\nThe implementation of ASTER reuses code from [Tensorflow Object Detection API](https://github.com/tensorflow/models/tree/master/research/object_detection).\n\n## Update\n**[07/13/2019] A PyTorch [port](https://github.com/ayumiymk/aster.pytorch) has been made by [@ayumiymk](https://github.com/ayumiymk).**\n\n## Correction (10/22/2018)\n\nWe have identified a bug we accidentally made in the code that causes only part of SVT images being tested and results in higher results. The bug has been fixed in commit [a7e8613](https://github.com/bgshih/aster/commit/a7e8613d6308e5a7aacb1237dfa0286d73cef342). Below are the corrected numbers on SVT. The results are still state-of-the-art, so the conclusions are not affected.\n\n  - SVT (50) ASTER: 97.4%; ASTER-A: 96.3%; ASTER-B: 96.1%; \n  - SVT (None): ASTER: 89.5%; ASTER-A: 80.2%; ASTER-B: 81.6%\n\n\n## Prerequisites\n\nASTER was developed and tested with **TensorFlow r1.4**. Higher versions may not work.\n\nASTER requires [Protocol Buffers](https://github.com/google/protobuf) (version\u003e=2.6). Besides, in Ubuntu 16.04:\n```\nsudo apt install cmake libcupti-dev\npip3 install --user protobuf tqdm numpy editdistance\n```\n\n## Installation\n  1. Go to `c_ops/` and run `build.sh` to build the custom operators\n  2. Execute `protoc aster/protos/*.proto --python_out=.` to build the protobuf files\n  3. Add `/path/to/aster` to `PYTHONPATH`, or set this variable for every run\n\n## Demo\n\nA demo program is located at `aster/demo.py`, accompanied with pretrained model files available on our [release page](https://github.com/bgshih/aster/releases). Download `model-demo.zip` and extract it under `aster/experiments/demo/` before running the demo.\n\nTo run the demo, simply execute:\n\n```\npython3 aster/demo.py\n```\n\nThis will output the recognition result of the demo image and the rectified image.\n\n## Training and on-the-fly evaluation\n\nData preparation scripts for several popular scene text datasets are located under `aster/tools`. See their source code for usage.\n\nTo run the example training, execute\n\n```\npython3 aster/train.py \\\n  --exp_dir experiments/demo \\\n  --num_clones 2\n```\n\nChange the configuration in `experiments/aster/trainval.prototxt` to configure your own training process.\n\nDuring the training, you can run a separate program to repeatedly evaluates the produced checkpoints.\n\n```\npython3 aster/eval.py \\\n   --exp_dir experiments/demo\n```\n\nEvaluation configuration is also in `trainval.prototxt`.\n\n## Citation\n\nIf you find this project helpful for your research, please cite the following papers:\n\n```\n@article{bshi2018aster,\n  author  = {Baoguang Shi and\n               Mingkun Yang and\n               Xinggang Wang and\n               Pengyuan Lyu and\n               Cong Yao and\n               Xiang Bai},\n  title   = {ASTER: An Attentional Scene Text Recognizer with Flexible Rectification},\n  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence}, \n  volume  = {}, \n  number  = {}, \n  pages   = {1-1},\n  year    = {2018}, \n}\n\n@inproceedings{ShiWLYB16,\n  author    = {Baoguang Shi and\n               Xinggang Wang and\n               Pengyuan Lyu and\n               Cong Yao and\n               Xiang Bai},\n  title     = {Robust Scene Text Recognition with Automatic Rectification},\n  booktitle = {2016 {IEEE} Conference on Computer Vision and Pattern Recognition,\n               {CVPR} 2016, Las Vegas, NV, USA, June 27-30, 2016},\n  pages     = {4168--4176},\n  year      = {2016}\n}\n```\n\nIMPORTANT NOTICE: Although this software is licensed under MIT, our intention is to make it free for academic research purposes. If you are going to use it in a product, we suggest you [contact us](xbai@hust.edu.cn) regarding possible patent issues.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbgshih%2Faster","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbgshih%2Faster","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbgshih%2Faster/lists"}