NOTE: I'm not a original Author. This repository is forked from https://github.com/argman/EAST.
- Bumped Tensorflow version to 2.1.0
- Minor change in web page.
- Tested on Ubuntu 20.04 and Windows 10
This is a tensorflow re-implementation of EAST: An Efficient and Accurate Scene Text Detector. The features are summarized blow:
- Online demo
- http://east.zxytim.com/
- Result example: http://east.zxytim.com/?r=48e5020a-7b7f-11e7-b776-f23c91e0703e
- CAVEAT: There's only one cpu core on the demo server. Simultaneous access will degrade response time.
- Only RBOX part is implemented.
- A fast Locality-Aware NMS in C++ provided by the paper's author.
- The pre-trained model provided achieves 80.83 F1-score on ICDAR 2015 Incidental Scene Text Detection Challenge using only training images from ICDAR 2015 and 2013. see here for the detailed results.
- Differences from original paper
- Use ResNet-50 rather than PVANET
- Use dice loss (optimize IoU of segmentation) rather than balanced cross entropy
- Use linear learning rate decay rather than staged learning rate decay
- Speed on 720p (resolution of 1280x720) images:
- Now
- Graphic card: GTX 1080 Ti
- Network fprop: ~50 ms
- NMS (C++): ~6ms
- Overall: ~16 fps
- Then
- Graphic card: K40
- Network fprop: ~150 ms
- NMS (python): ~300ms
- Overall: ~2 fps
- Now
Thanks for the author's (@zxytim) help! Please cite his paper if you find this useful.
- Any version of tensorflow version >= 2.1.0 should be ok.
- Models trained on ICDAR 2013 (training set) + ICDAR 2015 (training set): BaiduYun link GoogleDrive
- Resnet V1 50 provided by tensorflow slim: slim resnet v1 50
If you want to train the model, you should provide the dataset path, in the dataset path, a separate gt text file should be provided for each image and run
python multigpu_train.py --gpu_list=0 --input_size=512 --batch_size_per_gpu=14 --checkpoint_path=/tmp/east_icdar2015_resnet_v1_50_rbox/ \
--text_scale=512 --training_data_path=/data/ocr/icdar2015/ --geometry=RBOX --learning_rate=0.0001 --num_readers=24 \
--pretrained_model_path=/tmp/resnet_v1_50.ckpt
If you have more than one gpu, you can pass gpu ids to gpu_list(like --gpu_list=0,1,2,3)
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. See the examples in training_samples/
If you've downloaded the pre-trained model, you can setup a demo server by
python3 run_demo_server.py --checkpoint-path /tmp/east_icdar2015_resnet_v1_50_rbox/
Then open http://localhost:8769 for the web demo. Notice that the URL will change after you submitted an image.
Something like ?r=49647854-7ac2-11e7-8bb7-80000210fe80
appends and that makes the URL persistent.
As long as you are not deleting data in static/results
, you can share your results to your friends using
the same URL.
URL for example below: http://east.zxytim.com/?r=48e5020a-7b7f-11e7-b776-f23c91e0703e
run
python eval.py --test_data_path=/tmp/images/ --gpu_list=0 --checkpoint_path=/tmp/east_icdar2015_resnet_v1_50_rbox/ \
--output_dir=/tmp/
a text file will be then written to the output path.
Here are some test examples on icdar2015, enjoy the beautiful text boxes!
- How to compile lanms on Windows ? Environment is: windows10 , python3.6 (ps:I use Anaconda3), vs2015 Update 3 (ps: Visual Studio version should be vs2015 Update 3 or newer)
Step 1:Open the console make sure you can use X64 compiler:cl.exe(note: not X86, not X86, not X86 )
Step 2:to compile adaptor.cpp to adaptor.pyd
execute:cd your_dir\lanms
execute:cl adaptor.cpp ./include/clipper/clipper.cpp /I ./include /I "C:\ProgramData\Anaconda3\include" /LD /Fe:adaptor.pyd /link/LIBPATH:"C:\ProgramData\Anaconda3\libs"
ok,adaptor.pyd appears!
Step 3:comment some code in init.py
comment line7 and line8
Step 4:Congratulations! You can import lanms.