{"id":13543241,"url":"https://github.com/songdejia/EAST","last_synced_at":"2025-04-02T12:31:54.818Z","repository":{"id":148709767,"uuid":"146561919","full_name":"songdejia/EAST","owner":"songdejia","description":"This is a pytorch re-implementation of EAST: An Efficient and Accurate Scene Text Detector.","archived":false,"fork":false,"pushed_at":"2019-04-06T06:58:22.000Z","size":8531,"stargazers_count":577,"open_issues_count":31,"forks_count":153,"subscribers_count":13,"default_branch":"master","last_synced_at":"2024-11-03T10:32:10.602Z","etag":null,"topics":["deeplearning","east","icdar","ocr","pytorch","textdetection"],"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/songdejia.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}},"created_at":"2018-08-29T07:26:05.000Z","updated_at":"2024-10-08T08:38:30.000Z","dependencies_parsed_at":"2023-05-28T17:30:10.173Z","dependency_job_id":null,"html_url":"https://github.com/songdejia/EAST","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/songdejia%2FEAST","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/songdejia%2FEAST/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/songdejia%2FEAST/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/songdejia%2FEAST/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/songdejia","download_url":"https://codeload.github.com/songdejia/EAST/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":246815746,"owners_count":20838498,"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":["deeplearning","east","icdar","ocr","pytorch","textdetection"],"created_at":"2024-08-01T11:00:27.288Z","updated_at":"2025-04-02T12:31:49.808Z","avatar_url":"https://github.com/songdejia.png","language":"Python","funding_links":[],"categories":["Text detection and localization"],"sub_categories":["EAST [paper:2017](https://arxiv.org/abs/1704.03155)"],"readme":"# EAST: An Efficient and Accurate Scene Text Detector\n### Description:\nThis version will be updated soon, please pay attention to this work.\nThe motivation of this version is to build a easy-training model. \nThis version can automatically update best_model by comparing current hmean and the former.\nAt the same time, we can see evaluation info about every sample easily.\n\n+ 1.train\n+ 2.predict \n+ 3.compress\n+ 4.compute Hmean(if Hmean is higher than before, update best_weight.pkl)\n+ 5.visualization(blue, green, red)\n+ 6.multi-scale test (update soon)\n    multi-scale vis. (vis with score, scales)\n\n### Thanks\nThe version is ported from [argman/EAST](https://github.com/argman/EAST), from Tensorflow to Pytorch\n\n### Check On Website\nIf you have no confidence of the result of our program, you could use submit.zip to submit on [website](http://rrc.cvc.uab.es/?ch=2\u0026com=mymethods\u0026task=1),then you can see result of every image.\n\n### Performance\n+ right -- green || wrong -- red || miss -- blue\n![visualization](https://github.com/songdejia/east-pytorch/blob/master/screenshots/vis01.png)\n![visualization](https://github.com/songdejia/east-pytorch/blob/master/screenshots/vis02.png)\n\n\n\n+ recall/precision/hmean for every test image\n![hmean](https://github.com/songdejia/east-pytorch/blob/master/screenshots/hmean.png)\n\n### Introduction\nThis is a pytorch re-implementation of [EAST: An Efficient and Accurate Scene Text Detector](https://arxiv.org/abs/1704.03155v2).\nThe features are summarized blow:\n\n+ Only **RBOX** part is implemented.\n+ A fast Locality-Aware NMS in C++ provided by the paper's author.(g++/gcc version 6.0 + will be ok)\n+ Evalution see [here](http://rrc.cvc.uab.es/?ch=4\u0026com=evaluation\u0026view=method_samples\u0026task=1\u0026m=29855\u0026gtv=1) for the detailed results.\n+ Differences from original paper\n\t+ Use ResNet-50 rather than PVANET\n\t+ Use dice loss (optimize IoU of segmentation) rather than balanced cross entropy\n\t+ Use linear learning rate decay rather than staged learning rate decay\n\t\nThanks for the author's ([@zxytim](https://github.com/zxytim)) help!\nPlease cite his [paper](https://arxiv.org/abs/1704.03155v2) if you find this useful.\n\n### Contents\n1. [Installation](#installation)\n2. [Download](#download)\n3. [Prepare dataset/pretrain](#dataset)\n4. [Test](#train)\n5. [Train](#test)\n6. [Examples](#examples)\n\n\n### Installation\n1. Any version of pytorch version \u003e 0.4.0 should be ok.\n\n### Download\n1. Pretrained model is not provided temporarily. Web site is updating now, please continue to pay attention \n\n### Prepare dataset/pretrain weight\n[1]. dataset(you need to prepare for dataset for train and test)\nsuggestions: you could do a soft-link to root_to_this_program/dataset/train/img/*.jpg\n+ -- train  ./dataset/train/img/img_###.jpg \n\t    ./dataset/train/gt/img_###.txt (you need to change name)\n+ -- test   ./data/test/img_###.jpg (img only)\n+ -- gt.zip ./result/gt.zip(ICDAR15 gt.zip is avaliable on [website](http://rrc.cvc.uab.es/?ch=2\u0026com=mymethods\u0026task=1)\n\n** Note: you can download dataset here\n+ -- [ICDAR15](http://rrc.cvc.uab.es/?ch=4\u0026com=downloads)\n+ -- [ICDAR13](http://rrc.cvc.uab.es/?ch=2\u0026com=downloads)\n\n[2]. pretrained  \n+ In config.py set resume True and set checkpoint path/to/weight/file\n+ I will provide pretrianed weight soon\n\n[3]. check GPUs and CPUs \nyou can use following to check aviliable gpu, this is for train\n```\nwatch -n 0.1 nvidia-smi\n```\nthen, you will see 2,3 is avaliable, modify config.py\ngpu_ids = [0,1], gpu = 2, and modify run.sh - CUDA_VISIBLE_DEVICES=2,3\n\n\n\n### Train\nIf you want to train the model, you should provide the dataset path in config.py and run\n\n```\nsh run.py\n```\n** Note: you should modify run.sh to specify your gpu id\n\nIf you have more than one gpu, you can pass gpu ids to gpu_list(like gpu_list=0,1,2,3) in config.py\n\n** Note: you should change the gt text file of icdar2015's filename to img_\\*.txt instead of gt_img_\\*.txt(or you can change the code in icdar.py), and some extra characters should be removed from the file.\nSee the examples in training_samples/**\n\n### Test\nBy default, we set train-eval process into integer.\nIf you want to use eval independently, you can do it by yourself. Any question can contact me.\n\n\n### Examples\nHere are some test examples on icdar2015, enjoy the beautiful text boxes!\n![image_1](demo_images/img_2.jpg)\n![image_2](demo_images/img_10.jpg)\n![image_3](demo_images/img_14.jpg)\n![image_4](demo_images/img_26.jpg)\n![image_5](demo_images/img_75.jpg)\n\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsongdejia%2FEAST","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsongdejia%2FEAST","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsongdejia%2FEAST/lists"}