@@ -101,7 +101,7 @@ Considering that the features of some channels will be suppressed if the convolu
The recognition module of PP-OCRv3 is optimized based on the text recognition algorithm [SVTR](https://arxiv.org/abs/2205.00159). RNN is abandoned in SVTR, and the context information of the text line image is more effectively mined by introducing the Transformers structure, thereby improving the text recognition ability.
The recognition accuracy of SVTR_inty outperforms PP-OCRv2 recognition model by 5.3%, while the prediction speed nearly 11 times slower. It takes nearly 100ms to predict a text line on CPU. Therefore, as shown in the figure below, PP-OCRv3 adopts the following six optimization strategies to accelerate the recognition model.
The recognition accuracy of SVTR_tiny outperforms PP-OCRv2 recognition model by 5.3%, while the prediction speed nearly 11 times slower. It takes nearly 100ms to predict a text line on CPU. Therefore, as shown in the figure below, PP-OCRv3 adopts the following six optimization strategies to accelerate the recognition model.
@@ -200,7 +200,7 @@ UDML (Unified-Deep Mutual Learning) is a strategy proposed in PP-OCRv2 which is
**（6）UIM：Unlabeled Images Mining**
UIM (Unlabeled Images Mining) is a very simple unlabeled data mining strategy. The main idea is to use a high-precision text recognition model to predict unlabeled images to obtain pseudo-labels, and select samples with high prediction confidence as training data for training lightweight models. Using this strategy, the accuracy of the recognition model is further improved to 79.4% (+1%).
UIM (Unlabeled Images Mining) is a very simple unlabeled data mining strategy. The main idea is to use a high-precision text recognition model to predict unlabeled images to obtain pseudo-labels, and select samples with high prediction confidence as training data for training lightweight models. Using this strategy, the accuracy of the recognition model is further improved to 79.4% (+1%). In practice, we use the full data set to train the high-precision SVTR_Tiny model (acc=82.5%) for data mining. [SVTR_Tiny model download and tutorial](../../applications/高精度中文识别模型.md).
For English recognition model inference, you can execute the following commands,you need to specify the dictionary path used by `--rec_char_dict_path`:
After executing the command, the prediction result of the above figure is:
```
Predicts of ./doc/imgs_words/en/word_1.png: ('JOINT', 0.998160719871521)
```
<aname="3-multilingual-model-inference"></a>
### 3. Multilingual Model Inference
If you need to predict [other language models](./models_list_en.md#Multilingual), when using inference model prediction, you need to specify the dictionary path used by `--rec_char_dict_path`. At the same time, in order to get the correct visualization results,
You need to specify the visual font path through `--vis_font_path`. There are small language fonts provided by default under the `doc/fonts` path, such as Korean recognition:
@@ -29,10 +29,10 @@ PP-OCR pipeline is as follows:
PP-OCR system is in continuous optimization. At present, PP-OCR and PP-OCRv2 have been released:
PP-OCR 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 (as shown in the green box above). 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).
PP-OCR 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 (as shown in the green box above). 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 report](https://arxiv.org/abs/2009.09941).
#### PP-OCRv2
On the basis of PP-OCR, PP-OCRv2 is further optimized in five aspects. The detection model adopts CML(Collaborative Mutual Learning) knowledge distillation strategy and CopyPaste data expansion strategy. The recognition model adopts LCNet lightweight backbone network, U-DML knowledge distillation strategy and enhanced CTC loss function improvement (as shown in the red box above), which further improves the inference speed and prediction effect. For more details, please refer to the technical report of PP-OCRv2 (https://arxiv.org/abs/2109.03144).
On the basis of PP-OCR, PP-OCRv2 is further optimized in five aspects. The detection model adopts CML(Collaborative Mutual Learning) knowledge distillation strategy and CopyPaste data expansion strategy. The recognition model adopts LCNet lightweight backbone network, U-DML knowledge distillation strategy and enhanced CTC loss function improvement (as shown in the red box above), which further improves the inference speed and prediction effect. For more details, please refer to the [PP-OCRv2 technical report](https://arxiv.org/abs/2109.03144).
#### PP-OCRv3
...
...
@@ -46,7 +46,7 @@ PP-OCRv3 pipeline is as follows:
<imgsrc="../ppocrv3_framework.png"width="800">
</div>
For more details, please refer to [PP-OCRv3 technical report](./PP-OCRv3_introduction_en.md).
For more details, please refer to [PP-OCRv3 technical report](https://arxiv.org/abs/2206.03001v2).
@@ -446,7 +446,7 @@ class PaddleOCR(predict_system.TextSystem):
"""
ocr with paddleocr
args：
img: img for ocr, support ndarray, img_path and list or ndarray
img: img for ocr, support ndarray, img_path and list of ndarray
det: use text detection or not. If false, only rec will be exec. Default is True
rec: use text recognition or not. If false, only det will be exec. Default is True
cls: use angle classifier or not. Default is True. If true, the text with rotation of 180 degrees can be recognized. If no text is rotated by 180 degrees, use cls=False to get better performance. Text with rotation of 90 or 270 degrees can be recognized even if cls=False.
