revert dict and add space in character

README.md

@@ -5,6 +5,7 @@ PaddleOCR旨在打造一套丰富、领先、且实用的OCR工具库，助力
 
 **近期更新**
 - 2020.7.9 添加支持空格的识别模型，[识别效果](#支持空格的中文OCR效果展示)
+- 2020.7.9 添加数据增强、学习率衰减策略,具体参考[配置文件](./doc/doc_ch/config.md)
 - 2020.6.8 添加[数据集](./doc/doc_ch/datasets.md)，并保持持续更新
 - 2020.6.5 支持 `attetnion` 模型导出 `inference_model`
 - 2020.6.5 支持单独预测识别时，输出结果得分
@@ -51,6 +52,8 @@ mkdir inference && cd inference
 wget https://paddleocr.bj.bcebos.com/ch_models/ch_det_mv3_db_infer.tar && tar xf ch_det_mv3_db_infer.tar
 # 下载超轻量级中文OCR模型的识别模型并解压
 wget https://paddleocr.bj.bcebos.com/ch_models/ch_rec_mv3_crnn_infer.tar && tar xf ch_rec_mv3_crnn_infer.tar
+# 下载支持空格的超轻量级中文OCR模型的识别模型并解压
+wget https://paddleocr.bj.bcebos.com/ch_models/ch_rec_mv3_crnn_enhance_infer.tar && tar xf ch_rec_mv3_crnn_enhance_infer.tar
 cd ..
 ```
 #### (2)通用中文OCR模型下载
@@ -60,6 +63,8 @@ mkdir inference && cd inference
 wget https://paddleocr.bj.bcebos.com/ch_models/ch_det_r50_vd_db_infer.tar && tar xf ch_det_r50_vd_db_infer.tar
 # 下载通用中文OCR模型的识别模型并解压
 wget https://paddleocr.bj.bcebos.com/ch_models/ch_rec_r34_vd_crnn_infer.tar && tar xf ch_rec_r34_vd_crnn_infer.tar
+# 下载支持空格的通用中文OCR模型的识别模型并解压
+wget https://paddleocr.bj.bcebos.com/ch_models/ch_rec_r34_vd_crnn_enhance_infer.tar && tar xf ch_rec_r34_vd_crnn_enhance_infer.tar
 cd ..
 ```
 
@@ -85,6 +90,13 @@ python3 tools/infer/predict_system.py --image_dir="./doc/imgs/11.jpg" --det_mode
 python3 tools/infer/predict_system.py --image_dir="./doc/imgs/11.jpg" --det_model_dir="./inference/ch_det_r50_vd_db/"  --rec_model_dir="./inference/ch_rec_r34_vd_crnn/"
 ```
 
+带空格的通用中文OCR模型的体验可以按照上述步骤下载相应的模型，并且更新相关的参数，示例如下：
+
+```
+# 预测image_dir指定的单张图像
+python3 tools/infer/predict_system.py --image_dir="./doc/imgs_en/img_12.jpg" --det_model_dir="./inference/ch_det_r50_vd_db/"  --rec_model_dir="./inference/ch_rec_r34_vd_crnn_enhance/"
+```
+
 更多的文本检测、识别串联推理使用方式请参考文档教程中[基于预测引擎推理](./doc/doc_ch/inference.md)。
 
 ## 文档教程

README_en.md

@@ -5,6 +5,7 @@ PaddleOCR aims to create a rich, leading, and practical OCR tools that help user
 
 **Recent updates**、
 - 2020.7.9 Add recognition model to support space, [recognition result](#space Chinese OCR results)
+- 2020.7.9 Add data auguments and learning rate decay strategies,please read [config](./doc/doc_en/config_en.md)
 - 2020.6.8 Add [dataset](./doc/doc_en/datasets_en.md) and keep updating
 - 2020.6.5 Support exporting `attention` model to `inference_model`
 - 2020.6.5 Support separate prediction and recognition, output result score
@@ -52,6 +53,9 @@ mkdir inference && cd inference
 wget https://paddleocr.bj.bcebos.com/ch_models/ch_det_mv3_db_infer.tar && tar xf ch_det_mv3_db_infer.tar
 # Download the recognition part of the lightweight Chinese OCR and decompress it
 wget https://paddleocr.bj.bcebos.com/ch_models/ch_rec_mv3_crnn_infer.tar && tar xf ch_rec_mv3_crnn_infer.tar
+# Download the space-recognized part of the lightweight Chinese OCR and decompress it
+wget https://paddleocr.bj.bcebos.com/ch_models/ch_rec_mv3_crnn_enhance_infer.tar && tar xf ch_rec_mv3_crnn_enhance_infer.tar
+
 cd ..
 ```
 #### (2) Download General Chinese OCR models
@@ -61,6 +65,8 @@ mkdir inference && cd inference
 wget https://paddleocr.bj.bcebos.com/ch_models/ch_det_r50_vd_db_infer.tar && tar xf ch_det_r50_vd_db_infer.tar
 # Download the recognition part of the generic Chinese OCR model and decompress it
 wget https://paddleocr.bj.bcebos.com/ch_models/ch_rec_r34_vd_crnn_infer.tar && tar xf ch_rec_r34_vd_crnn_infer.tar
+# Download the space-recognition part of the generic Chinese OCR model and decompress it
+wget https://paddleocr.bj.bcebos.com/ch_models/ch_rec_r34_vd_crnn_enhance_infer.tar && tar xf ch_rec_r34_vd_crnn_enhance_infer.tar
 cd ..
 ```
 
@@ -86,6 +92,13 @@ To run inference of the Generic Chinese OCR model, follow these steps above to d
 python3 tools/infer/predict_system.py --image_dir="./doc/imgs/11.jpg" --det_model_dir="./inference/ch_det_r50_vd_db/"  --rec_model_dir="./inference/ch_rec_r34_vd_crnn/"
 ```
 
+To run inference of the space-Generic Chinese OCR model, follow these steps above to download the corresponding models and update the relevant parameters. Examples are as follows:
+
+```
+# Prediction on a single image by specifying image path to image_dir
+python3 tools/infer/predict_system.py --image_dir="./doc/imgs_en/img_12.jpg" --det_model_dir="./inference/ch_det_r50_vd_db/"  --rec_model_dir="./inference/ch_rec_r34_vd_crnn_enhance/"
+```
+
 For more text detection and recognition models, please refer to the document [Inference](./doc/doc_en/inference_en.md)
 
 ## DOCUMENTATION

configs/rec/rec_chinese_common_train.yml

@@ -14,6 +14,8 @@ Global:
   character_type: ch
   character_dict_path: ./ppocr/utils/ppocr_keys_v1.txt
   loss_type: ctc
+  distort: false
+  add_sapce: false
   reader_yml: ./configs/rec/rec_chinese_reader.yml
   pretrain_weights:
   checkpoints:

configs/rec/rec_chinese_lite_train.yml

@@ -15,6 +15,7 @@ Global:
   character_dict_path: ./ppocr/utils/ppocr_keys_v1.txt
   loss_type: ctc
   distort: false
+  add_space: false
   reader_yml: ./configs/rec/rec_chinese_reader.yml
   pretrain_weights:
   checkpoints:

doc/doc_ch/config.md

@@ -30,6 +30,8 @@
 |      character_type      |    设置字符类型            |       ch          |    en/ch, en时将使用默认dict，ch时使用自定义dict|
 |      character_dict_path |    设置字典路径            |  ./ppocr/utils/ic15_dict.txt  |    \                 |
 |      loss_type           |    设置 loss 类型              |       ctc         |    支持两种loss： ctc / attention |
+|       distort            |    设置是否使用数据增强          |       false       |  设置为true时，将在训练时随机进行扰动，支持的扰动操作可阅读[img_tools.py](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/ppocr/data/rec/img_tools.py)                 |
+|       add_space          |    设置是否识别空格             |        false      |          仅在"ch"模式下支持空格                 |
 |      reader_yml          |    设置reader配置文件          |  ./configs/rec/rec_icdar15_reader.yml  |  \          |
 |      pretrain_weights    |    加载预训练模型路径      |  ./pretrain_models/CRNN/best_accuracy  |  \          |
 |      checkpoints         |    加载模型参数路径            |       None        |    用于中断后加载参数继续训练 |

doc/doc_ch/recognition.md

@@ -94,7 +94,7 @@ word_dict.txt 每行有一个单字，将字符与数字索引映射在一起，
 `ppocr/utils/ic15_dict.txt` 是一个包含36个字符的英文字典，
 您可以按需使用。
 
-如需自定义dic文件，请修改 `configs/rec/rec_icdar15_train.yml` 中的 `character_dict_path` 字段, 并将 `character_type` 设置为 `ch`。
+如需自定义dic文件，请在 `configs/rec/rec_icdar15_train.yml` 中添加 `character_dict_path` 字段, 并将 `character_type` 设置为 `ch`。
 
 ### 启动训练
 
@@ -157,12 +157,26 @@ Global:
   character_type: ch
   # 添加自定义字典，如修改字典请将路径指向新字典
   character_dict_path: ./ppocr/utils/ppocr_keys_v1.txt
+  # 训练时添加数据增强
+  distort: true
+  # 识别空格
+  add_space: true
   ...
   # 修改reader类型
   reader_yml: ./configs/rec/rec_chinese_reader.yml
   ...
 
 ...
+
+Optimizer:
+  ...
+  # 添加学习率衰减策略
+  decay:
+    function: cosine_decay
+    # 每个 epoch 包含 iter 数
+    step_each_epoch: 20
+    # 总共训练epoch数
+    total_epoch: 1000
 ```
 **注意，预测/评估时的配置文件请务必与训练一致。**
 

doc/doc_en/config_en.md

@@ -30,6 +30,8 @@ Take `rec_chinese_lite_train.yml` as an example
 |      character_type      |    Set character type            |       ch          |    en/ch, the default dict will be used for en, and the custom dict will be used for ch|
 |      character_dict_path |    Set dictionary path            |  ./ppocr/utils/ic15_dict.txt  |    \                 |
 |      loss_type           |    Set loss type              |       ctc         |    Supports two types of loss: ctc / attention |
+|       distort            |    Set use distort          |       false       |  Support distort type ,read [img_tools.py](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/ppocr/data/rec/img_tools.py)                 |
+|       add_space          |    Wether to recognize space             |        false      |         Only support in "ch" mode                 |
 |      reader_yml          |    Set the reader configuration file          |  ./configs/rec/rec_icdar15_reader.yml  |  \          |
 |      pretrain_weights    |    Load pre-trained model path      |  ./pretrain_models/CRNN/best_accuracy  |  \          |
 |      checkpoints         |    Load saved model path            |       None        |    Used to load saved parameters to continue training after interruption |

doc/doc_en/recognition_en.md

@@ -158,9 +158,23 @@ Global:
   ...
   # Modify reader type
   reader_yml: ./configs/rec/rec_chinese_reader.yml
+  # Whether to use data augmentation
+  distort: true
+  # Whether to recognize spaces
+  add_space: true
   ...
 
 ...
+
+Optimizer:
+  ...
+  # Add learning rate decay strategy
+  decay:
+    function: cosine_decay
+    # Each epoch contains iter number
+    step_each_epoch: 20
+    # Total epoch number
+    total_epoch: 1000
 ```
 **Note that the configuration file for prediction/evaluation must be consistent with the training.**
 

ppocr/data/rec/dataset_traversal.py

@@ -58,6 +58,7 @@ class LMDBReader(object):
         else:
             self.batch_size = params['test_batch_size_per_card']
             self.drop_last = False
+            self.use_distort = False
         self.infer_img = params['infer_img']
 
     def load_hierarchical_lmdb_dataset(self):
@@ -206,6 +207,7 @@ class SimpleReader(object):
         else:
             self.batch_size = params['test_batch_size_per_card']
             self.drop_last = False
+            self.use_distort = False
 
     def __call__(self, process_id):
         if self.mode != 'train':

ppocr/data/rec/img_tools.py

@@ -136,6 +136,9 @@ def jitter(img):
 
 
 def add_gasuss_noise(image, mean=0, var=0.1):
+    """
+    Gasuss noise
+    """
 
     noise = np.random.normal(mean, var**0.5, image.shape)
     out = image + 0.5 * noise
@@ -152,9 +155,8 @@ def get_crop(image):
     top_min = 1
     top_max = 8
     top_crop = int(random.randint(top_min, top_max))
-
+    top_crop = min(top_crop, h - 1)
     crop_img = image.copy()
-
     ratio = random.randint(0, 1)
     if ratio:
         crop_img = crop_img[top_crop:h, :, :]
@@ -249,13 +251,13 @@ def get_warpR(config):
     dst2 = r.dot(p2)
     dst3 = r.dot(p3)
     dst4 = r.dot(p4)
-    list_dst = [dst1, dst2, dst3, dst4]
+    list_dst = np.array([dst1, dst2, dst3, dst4])
     org = np.array([[0, 0], [w, 0], [0, h], [w, h]], np.float32)
     dst = np.zeros((4, 2), np.float32)
     # Project onto the image plane
-    for i in range(4):
-        dst[i, 0] = list_dst[i][0] * z / (z - list_dst[i][2]) + pcenter[0]
-        dst[i, 1] = list_dst[i][1] * z / (z - list_dst[i][2]) + pcenter[1]
+    dst[:, 0] = list_dst[:, 0] * z / (z - list_dst[:, 2]) + pcenter[0]
+    dst[:, 1] = list_dst[:, 1] * z / (z - list_dst[:, 2]) + pcenter[1]
+
     warpR = cv2.getPerspectiveTransform(org, dst)
 
     dst1, dst2, dst3, dst4 = dst

ppocr/utils/character.py

@@ -30,12 +30,17 @@ class CharacterOps(object):
             dict_character = list(self.character_str)
         elif self.character_type == "ch":
             character_dict_path = config['character_dict_path']
+            add_space = False
+            if 'add_space' in config:
+                add_space = config['add_space']
             self.character_str = ""
             with open(character_dict_path, "rb") as fin:
                 lines = fin.readlines()
                 for line in lines:
                     line = line.decode('utf-8').strip("\n").strip("\r\n")
                     self.character_str += line
+            if add_space:
+                self.character_str += " "
             dict_character = list(self.character_str)
         elif self.character_type == "en_sensitive":
             # same with ASTER setting (use 94 char).

ppocr/utils/ppocr_keys_v1.txt

@@ -6620,5 +6620,4 @@ j
 緖
 續
 紹
-懮
- 
+懮
\ No newline at end of file

tools/infer/predict_rec.py

@@ -39,7 +39,8 @@ class TextRecognizer(object):
         self.rec_algorithm = args.rec_algorithm
         char_ops_params = {
             "character_type": args.rec_char_type,
-            "character_dict_path": args.rec_char_dict_path
+            "character_dict_path": args.rec_char_dict_path,
+            "add_space": args.rec_add_space
         }
         if self.rec_algorithm != "RARE":
             char_ops_params['loss_type'] = 'ctc'

tools/infer/utility.py

@@ -63,6 +63,7 @@ def parse_args():
         "--rec_char_dict_path",
         type=str,
         default="./ppocr/utils/ppocr_keys_v1.txt")
+    parser.add_argument("--rec_add_space", type=bool, default=True)
     return parser.parse_args()