PaddleOCR aims to create rich, leading, and practical OCR tools that help users train better models and apply them into practice.
**Recent updates**
- 2020.9.19 Update the ultra lightweight compressed ppocr_mobile_slim series models, the overall model size is 3.5M (see [PP-OCR Pipline](#PP-OCR)), suitable for mobile deployment. [Model Downloads](#Supported-Chinese-model-list)
- 2020.9.22 Update the PP-OCR technical article, https://arxiv.org/abs/2009.09941
- 2020.9.19 Update the ultra lightweight compressed ppocr_mobile_slim series models, the overall model size is 3.5M (see [PP-OCR Pipeline](#PP-OCR-Pipeline)), suitable for mobile deployment. [Model Downloads](#Supported-Chinese-model-list)
- 2020.9.17 Update the ultra lightweight ppocr_mobile series and general ppocr_server series Chinese and English ocr models, which are comparable to commercial effects. [Model Downloads](#Supported-Chinese-model-list)
- 2020.8.24 Support the use of PaddleOCR through whl package installation，pelease refer [PaddleOCR Package](./doc/doc_en/whl_en.md)
- 2020.9.17 update [English recognition model](./doc/doc_en/models_list_en.md#english-recognition-model) and [Multilingual recognition model](doc/doc_en/models_list_en.md#english-recognition-model), `German`, `French`, `Japanese` and `Korean` have been supported. Models for more languages will continue to be updated.
- 2020.8.24 Support the use of PaddleOCR through whl package installation，please refer [PaddleOCR Package](./doc/doc_en/whl_en.md)
- 2020.8.21 Update the replay and PPT of the live lesson at Bilibili on August 18, lesson 2, easy to learn and use OCR tool spree. [Get Address](https://aistudio.baidu.com/aistudio/education/group/info/1519)
- [more](./doc/doc_en/update_en.md)
@ -29,7 +30,16 @@ PaddleOCR aims to create rich, leading, and practical OCR tools that help users
<imgsrc="doc/imgs_results/1103.jpg"width="800">
</div>
The above picture is the effect display of the general ppocr_server model. For more effect pictures, please see [More visualization](./doc/doc_en/visualization_en.md).
The above pictures are the visualizations of the general ppocr_server model. For more effect pictures, please see [More visualizations](./doc/doc_en/visualization_en.md).
<aname="Community"></a>
## Community
- Scan the QR code below with your Wechat, you can access to official technical exchange group. Look forward to your participation.
- [Quick inference based on pip](./doc/doc_en/whl_en.md)
@ -92,35 +103,40 @@ For more model downloads (including multiple languages), please refer to [PP-OCR
- [License](#LICENSE)
- [Contribution](#CONTRIBUTION)
<aname="PP-OCR-Pipline"></a>
<aname="PP-OCR-Pipeline"></a>
## PP-OCR Pipline
## PP-OCR Pipeline
<divalign="center">
<imgsrc="./doc/ppocr_framework.png"width="800">
</div>
PP-OCR is a practical ultra-lightweight OCR system. It is mainly composed of three parts: DB text detection, detection frame correction and CRNN text recognition. The system adopts 19 effective strategies from 8 aspects including backbone network selection and adjustment, prediction head design, data augmentation, learning rate transformation strategy, regularization parameter selection, pre-training model use, and automatic model tailoring and quantization to optimize and slim down the models of each module. The final results are an ultra-lightweight Chinese and English OCR model with an overall size of 3.5M and a 2.8M English digital OCR model. For more details, please refer to the PP-OCR technical article (Arxiv article link is being generated).
PP-OCR is a practical ultra-lightweight OCR system. It is mainly composed of three parts: DB text detection, detection frame correction and CRNN text recognition. The system adopts 19 effective strategies from 8 aspects including backbone network selection and adjustment, prediction head design, data augmentation, learning rate transformation strategy, regularization parameter selection, pre-training model use, and automatic model tailoring and quantization to optimize and slim down the models of each module. The final results are an ultra-lightweight Chinese and English OCR model with an overall size of 3.5M and a 2.8M English digital OCR model. For more details, please refer to the PP-OCR technical article (https://arxiv.org/abs/2009.09941). Besides, The implementation of the FPGM Pruner and PACT quantization is based on [PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim).
* If you want to orientation classifier to correct the detected boxes, you can set `use_angle_cls` in the file `tools/config.txt` as 1 to enable the function.
* What's more, Parameters and their meanings in `tools/config.txt` are as follows.
```
use_gpu 0 # Whether to use GPU, 0 means not to use, 1 means to use
gpu_id 0 # GPU id when use_gpu is 1
gpu_mem 4000 # GPU memory requested
cpu_math_library_num_threads 10 # Number of threads when using CPU inference. When machine cores is enough, the large the value, the faster the inference speed
use_mkldnn 1 # Whether to use mkdlnn library
use_zero_copy_run 1 # Whether to use use_zero_copy_run for inference
max_side_len 960 # Limit the maximum image height and width to 960
det_db_thresh 0.3 # Used to filter the binarized image of DB prediction, setting 0.-0.3 has no obvious effect on the result
det_db_box_thresh 0.5 # DDB post-processing filter box threshold, if there is a missing box detected, it can be reduced as appropriate
det_db_unclip_ratio 1.6 # Indicates the compactness of the text box, the smaller the value, the closer the text box to the text
det_model_dir ./inference/det_db # Address of detection inference model
# cls config
use_angle_cls 0 # Whether to use the direction classifier, 0 means not to use, 1 means to use
cls_model_dir ./inference/cls # Address of direction classifier inference model
cls_thresh 0.9 # Score threshold of the direction classifier
# rec config
rec_model_dir ./inference/rec_crnn # Address of recognition inference model
visualize 1 # Whether to visualize the results，when it is set as 1, The prediction result will be save in the image file `./ocr_vis.png`.
```
* Multi-language inference is also supported in PaddleOCR, for more details, please refer to part of multi-language dictionaries and models in [recognition tutorial](../../doc/doc_en/recognition_en.md).
The detection results will be shown on the screen, which is as follows.
@ -210,6 +242,6 @@ The detection results will be shown on the screen, which is as follows.
</div>
### 2.3 Note
### 2.3 Notes
* `MKLDNN` is disabled by default for C++ inference (`use_mkldnn` in `tools/config.txt` is set to 0), if you need to use MKLDNN for inference acceleration, you need to modify `use_mkldnn` to 1, and use the latest version of the Paddle source code to compile the inference library. When using MKLDNN for CPU prediction, if multiple images are predicted at the same time, there will be a memory leak problem (the problem is not present if MKLDNN is disabled). The problem is currently being fixed, and the temporary solution is: when predicting multiple pictures, Re-initialize the recognition (`CRNNRecognizer`) and detection class (`DBDetector`) every 30 pictures or so.
* Paddle2.0.0-beta0 inference model library is recommanded for this tuturial.
WenmuZhou
6 years ago