@@ -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 can be increased from 74.8% to 80.1% (+5.3%) by replacing the recognition model of PP-OCRv2 with SVTR_Tiny, 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_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.
@@ -177,7 +177,7 @@ Note: When testing the speed, the input image shape of 01-05 are all (3, 32, 320
**（3）TextConAug：Data Augmentation Strategy for Mining Text Context Information**
TextConAug is a data augmentation strategy for mining textual context information. The main idea comes from the paper [ConCLR](https://www.cse.cuhk.edu.hk/~byu/papers/C139-AAAI2022-ConCLR.pdf), in which the author proposes data augmentation strategy ConAug to connect 2 different images in a batch to form new images and perform self-supervised comparative learning. PP-OCRv3 applies this method to supervised learning tasks, and designs the TextConAug data augmentation method, which can enrich the context information of training data and improve the diversity of training data. Using this strategy, the accuracy of the recognition model is further improved to 76.3% (+0.5%). The schematic diagram of TextConAug is as follows:
TextConAug is a data augmentation strategy for mining textual context information. The main idea comes from the paper [ConCLR](https://www.cse.cuhk.edu.hk/~byu/papers/C139-AAAI2022-ConCLR.pdf), in which the author proposes data augmentation strategy ConAug to concat 2 different images in a batch to form new images and perform self-supervised comparative learning. PP-OCRv3 applies this method to supervised learning tasks, and designs the TextConAug data augmentation method, which can enrich the context information of training data and improve the diversity of training data. Using this strategy, the accuracy of the recognition model is further improved to 76.3% (+0.5%). The schematic diagram of TextConAug is as follows:
<divalign="center">
<imgsrc="../ppocr_v3/recconaug.png"width=800>
...
...
@@ -195,7 +195,7 @@ TextRotNet is a pre-trained model trained with a large amount of unlabeled text
**（5）UDML：Unified-Deep Mutual Learning**
UDML (Unified-Deep Mutual Learning) is a strategy proposed in PP-OCRv2 which is very effective to improve the model effect. In PP-OCRv3, for two different structures SVTR_LCNet and Attention, the feature map of PP-LCNet, the output of the SVTR module and the output of the Attention module between them are simultaneously supervised and trained. Using this strategy, the accuracy of the recognition model is further improved to 78.4% (+1.5%).
UDML (Unified-Deep Mutual Learning) is a strategy proposed in PP-OCRv2 which is very effective to improve the model accuracy. In PP-OCRv3, for two different structures SVTR_LCNet and Attention, the feature map of PP-LCNet, the output of the SVTR module and the output of the Attention module between them are simultaneously supervised and trained. Using this strategy, the accuracy of the recognition model is further improved to 78.4% (+1.5%).
