Merge pull request #1780 from WenmuZhou/dygraph_rc

add starnet

doc/doc_ch/algorithm_overview.md

@@ -40,17 +40,19 @@ PaddleOCR文本检测算法的训练和使用请参考文档教程中[模型训
 PaddleOCR基于动态图开源的文本识别算法列表：
 - [x]  CRNN([paper](https://arxiv.org/abs/1507.05717))[7]（ppocr推荐）
 - [x]  Rosetta([paper](https://arxiv.org/abs/1910.05085))[10]
-- [ ]  STAR-Net([paper](http://www.bmva.org/bmvc/2016/papers/paper043/index.html))[11] coming soon
+- [x]  STAR-Net([paper](http://www.bmva.org/bmvc/2016/papers/paper043/index.html))[11]
 - [ ]  RARE([paper](https://arxiv.org/abs/1603.03915v1))[12] coming soon
 - [ ]  SRN([paper](https://arxiv.org/abs/2003.12294))[5] coming soon
 
 参考[DTRB][3](https://arxiv.org/abs/1904.01906)文字识别训练和评估流程，使用MJSynth和SynthText两个文字识别数据集训练，在IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE数据集上进行评估，算法效果如下：
 
 |模型|骨干网络|Avg Accuracy|模型存储命名|下载链接|
-|-|-|-|-|-|
+|---|---|---|---|---|
 |Rosetta|Resnet34_vd|80.9%|rec_r34_vd_none_none_ctc|[下载链接](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_r34_vd_none_none_ctc_v2.0_train.tar)|
 |Rosetta|MobileNetV3|78.05%|rec_mv3_none_none_ctc|[下载链接](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_mv3_none_none_ctc_v2.0_train.tar)|
 |CRNN|Resnet34_vd|82.76%|rec_r34_vd_none_bilstm_ctc|[下载链接](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_r34_vd_none_bilstm_ctc_v2.0_train.tar)|
 |CRNN|MobileNetV3|79.97%|rec_mv3_none_bilstm_ctc|[下载链接](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_mv3_none_bilstm_ctc_v2.0_train.tar)|
+|StarNet|Resnet34_vd|84.44%|rec_r34_vd_tps_bilstm_ctc|[下载链接](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_r34_vd_tps_bilstm_ctc_v2.0_train.tar)|
+|StarNet|MobileNetV3|81.42%|rec_mv3_tps_bilstm_ctc|[下载链接](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_mv3_tps_bilstm_ctc_v2.0_train.tar)|
 
 PaddleOCR文本识别算法的训练和使用请参考文档教程中[模型训练/评估中的文本识别部分](./recognition.md)。

doc/doc_ch/inference.md

@@ -352,10 +352,10 @@ Predicts of ./doc/imgs_words/ch/word_4.jpg:['0', 0.9999982]
 
 ```
 # 使用方向分类器
-python3 tools/infer/predict_system.py --image_dir="./doc/imgs/2.jpg" --det_model_dir="./inference/det_db/" --cls_model_dir="./inference/cls/" --rec_model_dir="./inference/rec_crnn/" --use_angle_cls=true
+python3 tools/infer/predict_system.py --image_dir="./doc/imgs/00018069.jpg" --det_model_dir="./inference/det_db/" --cls_model_dir="./inference/cls/" --rec_model_dir="./inference/rec_crnn/" --use_angle_cls=true
 
 # 不使用方向分类器
-python3 tools/infer/predict_system.py --image_dir="./doc/imgs/2.jpg" --det_model_dir="./inference/det_db/" --rec_model_dir="./inference/rec_crnn/" --use_angle_cls=false
+python3 tools/infer/predict_system.py --image_dir="./doc/imgs/00018069.jpg" --det_model_dir="./inference/det_db/" --rec_model_dir="./inference/rec_crnn/" --use_angle_cls=false
 ```
 
 
@@ -364,7 +364,7 @@ python3 tools/infer/predict_system.py --image_dir="./doc/imgs/2.jpg" --det_model
 
 执行命令后，识别结果图像如下：
 
-![](../imgs_results/2.jpg)
+![](../imgs_results/system_res_00018069.jpg)
 
 <a name="其他模型推理"></a>
 ### 2. 其他模型推理
@@ -381,4 +381,4 @@ python3 tools/infer/predict_system.py --image_dir="./doc/imgs_en/img_10.jpg" --d
 
 执行命令后，识别结果图像如下：
 
-(coming soon)
+![](../imgs_results/img_10_east_starnet.jpg)

doc/doc_en/algorithm_overview_en.md

@@ -41,17 +41,19 @@ For the training guide and use of PaddleOCR text detection algorithms, please re
 PaddleOCR open-source text recognition algorithms list:
 - [x]  CRNN([paper](https://arxiv.org/abs/1507.05717))[7]
 - [x]  Rosetta([paper](https://arxiv.org/abs/1910.05085))[10]
-- [ ]  STAR-Net([paper](http://www.bmva.org/bmvc/2016/papers/paper043/index.html))[11] coming soon
+- [x]  STAR-Net([paper](http://www.bmva.org/bmvc/2016/papers/paper043/index.html))[11]
 - [ ]  RARE([paper](https://arxiv.org/abs/1603.03915v1))[12] coming soon
 - [ ]  SRN([paper](https://arxiv.org/abs/2003.12294))[5] coming soon
 
 Refer to [DTRB](https://arxiv.org/abs/1904.01906), the training and evaluation result of these above text recognition (using MJSynth and SynthText for training, evaluate on IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE) is as follow:
 
 |Model|Backbone|Avg Accuracy|Module combination|Download link|
-|-|-|-|-|-|
+|---|---|---|---|---|
 |Rosetta|Resnet34_vd|80.9%|rec_r34_vd_none_none_ctc|[Download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_r34_vd_none_none_ctc_v2.0_train.tar)|
 |Rosetta|MobileNetV3|78.05%|rec_mv3_none_none_ctc|[Download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_mv3_none_none_ctc_v2.0_train.tar)|
 |CRNN|Resnet34_vd|82.76%|rec_r34_vd_none_bilstm_ctc|[Download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_r34_vd_none_bilstm_ctc_v2.0_train.tar)|
 |CRNN|MobileNetV3|79.97%|rec_mv3_none_bilstm_ctc|[Download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_mv3_none_bilstm_ctc_v2.0_train.tar)|
+|StarNet|Resnet34_vd|84.44%|rec_r34_vd_tps_bilstm_ctc|[Download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_r34_vd_tps_bilstm_ctc_v2.0_train.tar)|
+|StarNet|MobileNetV3|81.42%|rec_mv3_tps_bilstm_ctc|[Download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_mv3_tps_bilstm_ctc_v2.0_train.tar)|
 
 Please refer to the document for training guide and use of PaddleOCR text recognition algorithms [Text recognition model training/evaluation/prediction](./recognition_en.md)

doc/doc_en/inference_en.md

@@ -366,15 +366,15 @@ When performing prediction, you need to specify the path of a single image or a
 
 ```
 # use direction classifier
-python3 tools/infer/predict_system.py --image_dir="./doc/imgs/2.jpg" --det_model_dir="./inference/det_db/" --cls_model_dir="./inference/cls/" --rec_model_dir="./inference/rec_crnn/" --use_angle_cls=true
+python3 tools/infer/predict_system.py --image_dir="./doc/imgs/00018069.jpg" --det_model_dir="./inference/det_db/" --cls_model_dir="./inference/cls/" --rec_model_dir="./inference/rec_crnn/" --use_angle_cls=true
 
 # not use use direction classifier
-python3 tools/infer/predict_system.py --image_dir="./doc/imgs/2.jpg" --det_model_dir="./inference/det_db/" --rec_model_dir="./inference/rec_crnn/"
+python3 tools/infer/predict_system.py --image_dir="./doc/imgs/00018069.jpg" --det_model_dir="./inference/det_db/" --rec_model_dir="./inference/rec_crnn/"
 ```
 
 After executing the command, the recognition result image is as follows:
 
-![](../imgs_results/2.jpg)
+![](../imgs_results/system_res_00018069.jpg)
 
 <a name="OTHER_MODELS"></a>
 ### 2. OTHER MODELS
@@ -391,4 +391,4 @@ python3 tools/infer/predict_system.py --image_dir="./doc/imgs_en/img_10.jpg" --d
 
 After executing the command, the recognition result image is as follows:
 
-(coming soon)
+![](../imgs_results/img_10_east_starnet.jpg)

ppocr/modeling/transforms/tps.py

@@ -213,16 +213,14 @@ class GridGenerator(nn.Layer):
 
     def build_P_paddle(self, I_r_size):
         I_r_height, I_r_width = I_r_size
-        I_r_grid_x = paddle.divide(
+        I_r_grid_x = (paddle.arange(
-            paddle.arange(
+            -I_r_width, I_r_width, 2, dtype='float64') + 1.0
-                -I_r_width, I_r_width, 2, dtype='float64') + 1.0,
+                      ) / paddle.to_tensor(np.array([I_r_width]))
-            paddle.to_tensor(
+
-                I_r_width, dtype='float64'))
+        I_r_grid_y = (paddle.arange(
-        I_r_grid_y = paddle.divide(
+            -I_r_height, I_r_height, 2, dtype='float64') + 1.0
-            paddle.arange(
+                      ) / paddle.to_tensor(np.array([I_r_height]))
-                -I_r_height, I_r_height, 2, dtype='float64') + 1.0,
+
-            paddle.to_tensor(
-                I_r_height, dtype='float64'))  # self.I_r_height
         # P: self.I_r_width x self.I_r_height x 2
         P = paddle.stack(paddle.meshgrid(I_r_grid_x, I_r_grid_y), axis=2)
         P = paddle.transpose(P, perm=[1, 0, 2])

ppocr/postprocess/rec_postprocess.py

@@ -109,7 +109,7 @@ class CTCLabelDecode(BaseRecLabelDecode):
 
         preds_idx = preds.argmax(axis=2)
         preds_prob = preds.max(axis=2)
-        text = self.decode(preds_idx, preds_prob)
+        text = self.decode(preds_idx, preds_prob, is_remove_duplicate=True)
         if label is None:
             return text
         label = self.decode(label)