recognition_en.md 14.4 KB
Newer Older
X
xxxpsyduck 已提交
1
## TEXT RECOGNITION
K
Khanh Tran 已提交
2

W
WenmuZhou 已提交
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
- [DATA PREPARATION](#DATA_PREPARATION)
    - [Dataset Download](#Dataset_download)
    - [Costom Dataset](#Costom_Dataset)  
    - [Dictionary](#Dictionary)  
    - [Add Space Category](#Add_space_category)

- [TRAINING](#TRAINING)
    - [Data Augmentation](#Data_Augmentation)
    - [Training](#Training)
    - [Multi-language](#Multi_language)

- [EVALUATION](#EVALUATION)

- [PREDICTION](#PREDICTION)
    - [Training engine prediction](#Training_engine_prediction)

<a name="DATA_PREPARATION"></a>
X
xxxpsyduck 已提交
20
### DATA PREPARATION
K
Khanh Tran 已提交
21 22 23 24 25 26 27 28 29


PaddleOCR supports two data formats: `LMDB` is used to train public data and evaluation algorithms; `general data` is used to train your own data:

Please organize the dataset as follows:

The default storage path for training data is `PaddleOCR/train_data`, if you already have a dataset on your disk, just create a soft link to the dataset directory:

```
30
ln -sf <path/to/dataset> <path/to/paddle_ocr>/train_data/dataset
K
Khanh Tran 已提交
31 32
```

W
WenmuZhou 已提交
33
<a name="Dataset_download"></a>
K
Khanh Tran 已提交
34 35 36 37
* Dataset download

If you do not have a dataset locally, you can download it on the official website [icdar2015](http://rrc.cvc.uab.es/?ch=4&com=downloads). Also refer to [DTRB](https://github.com/clovaai/deep-text-recognition-benchmark#download-lmdb-dataset-for-traininig-and-evaluation-from-here),download the lmdb format dataset required for benchmark

T
tink2123 已提交
38 39
If you want to reproduce the paper indicators of SRN, you need to download offline [augmented data](https://pan.baidu.com/s/1-HSZ-ZVdqBF2HaBZ5pRAKA), extraction code: y3ry. The augmented data is obtained by rotation and perturbation of mjsynth and synthtext. Please unzip the data to {your_path}/PaddleOCR/train_data/data_lmdb_Release/training/path.

W
WenmuZhou 已提交
40
<a name="Costom_Dataset"></a>
K
Khanh Tran 已提交
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92
* Use your own dataset:

If you want to use your own data for training, please refer to the following to organize your data.

- Training set

First put the training images in the same folder (train_images), and use a txt file (rec_gt_train.txt) to store the image path and label.

* Note: by default, the image path and image label are split with \t, if you use other methods to split, it will cause training error

```
" Image file name           Image annotation "

train_data/train_0001.jpg   简单可依赖
train_data/train_0002.jpg   用科技让复杂的世界更简单
```
PaddleOCR provides label files for training the icdar2015 dataset, which can be downloaded in the following ways:

```
# Training set label
wget -P ./train_data/ic15_data  https://paddleocr.bj.bcebos.com/dataset/rec_gt_train.txt
# Test Set Label
wget -P ./train_data/ic15_data  https://paddleocr.bj.bcebos.com/dataset/rec_gt_test.txt
```

The final training set should have the following file structure:

```
|-train_data
    |-ic15_data
        |- rec_gt_train.txt
        |- train
            |- word_001.png
            |- word_002.jpg
            |- word_003.jpg
            | ...
```

- Test set

Similar to the training set, the test set also needs to be provided a folder containing all images (test) and a rec_gt_test.txt. The structure of the test set is as follows:

```
|-train_data
    |-ic15_data
        |- rec_gt_test.txt
        |- test
            |- word_001.jpg
            |- word_002.jpg
            |- word_003.jpg
            | ...
```
W
WenmuZhou 已提交
93
<a name="Dictionary"></a>
K
Khanh Tran 已提交
94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
- Dictionary

Finally, a dictionary ({word_dict_name}.txt) needs to be provided so that when the model is trained, all the characters that appear can be mapped to the dictionary index.

Therefore, the dictionary needs to contain all the characters that you want to be recognized correctly. {word_dict_name}.txt needs to be written in the following format and saved in the `utf-8` encoding format:

```
l
d
a
d
r
n
```

In `word_dict.txt`, there is a single word in each line, which maps characters and numeric indexes together, e.g "and" will be mapped to [2 5 1]

`ppocr/utils/ppocr_keys_v1.txt` is a Chinese dictionary with 6623 characters.

W
WenmuZhou 已提交
113 114 115 116
`ppocr/utils/ic15_dict.txt` is an English dictionary with 63 characters

`ppocr/utils/dict/french_dict.txt` is a French dictionary with 118 characters

T
tink2123 已提交
117
`ppocr/utils/dict/japan_dict.txt` is a Japan dictionary with 4399 characters
W
WenmuZhou 已提交
118

T
tink2123 已提交
119
`ppocr/utils/dict/korean_dict.txt` is a Korean dictionary with 3636 characters
W
WenmuZhou 已提交
120

T
tink2123 已提交
121 122 123
`ppocr/utils/dict/german_dict.txt` is a German dictionary with 131 characters

`ppocr/utils/dict/en_dict.txt` is a English dictionary with 63 characters
W
WenmuZhou 已提交
124 125 126 127 128

You can use it on demand.

The current multi-language model is still in the demo stage and will continue to optimize the model and add languages. **You are very welcome to provide us with dictionaries and fonts in other languages**,
If you like, you can submit the dictionary file to [dict](../../ppocr/utils/dict) or corpus file to [corpus](../../ppocr/utils/corpus) and we will thank you in the Repo.
K
Khanh Tran 已提交
129 130 131 132


To customize the dict file, please modify the `character_dict_path` field in `configs/rec/rec_icdar15_train.yml` and set `character_type` to `ch`.

T
tink2123 已提交
133 134 135 136
- Custom dictionary

If you need to customize dic file, please add character_dict_path field in configs/rec/rec_icdar15_train.yml to point to your dictionary path. And set character_type to ch.

W
WenmuZhou 已提交
137
<a name="Add_space_category"></a>
T
tink2123 已提交
138 139
- Add space category

140
If you want to support the recognition of the `space` category, please set the `use_space_char` field in the yml file to `True`.
T
tink2123 已提交
141 142 143

**Note: use_space_char only takes effect when character_type=ch**

W
WenmuZhou 已提交
144
<a name="TRAINING"></a>
X
xxxpsyduck 已提交
145
### TRAINING
K
Khanh Tran 已提交
146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162

PaddleOCR provides training scripts, evaluation scripts, and prediction scripts. In this section, the CRNN recognition model will be used as an example:

First download the pretrain model, you can download the trained model to finetune on the icdar2015 data:

```
cd PaddleOCR/
# Download the pre-trained model of MobileNetV3
wget -P ./pretrain_models/ https://paddleocr.bj.bcebos.com/rec_mv3_none_bilstm_ctc.tar
# Decompress model parameters
cd pretrain_models
tar -xf rec_mv3_none_bilstm_ctc.tar && rm -rf rec_mv3_none_bilstm_ctc.tar
```

Start training:

```
163
# GPU training Support single card and multi-card training, specify the card number through --gpus
W
WenmuZhou 已提交
164
# Training icdar15 English data and saving the log as train_rec.log
165
python3 -m paddle.distributed.launch --gpus '0,1,2,3'  tools/train.py -c configs/rec/rec_icdar15_train.yml
K
Khanh Tran 已提交
166
```
W
WenmuZhou 已提交
167
<a name="Data_Augmentation"></a>
T
tink2123 已提交
168 169 170 171 172 173 174 175
- Data Augmentation

PaddleOCR provides a variety of data augmentation methods. If you want to add disturbance during training, please set `distort: true` in the configuration file.

The default perturbation methods are: cvtColor, blur, jitter, Gasuss noise, random crop, perspective, color reverse.

Each disturbance method is selected with a 50% probability during the training process. For specific code implementation, please refer to: [img_tools.py](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/ppocr/data/rec/img_tools.py)

W
WenmuZhou 已提交
176
<a name="Training"></a>
T
tink2123 已提交
177 178
- Training

K
Khanh Tran 已提交
179 180 181 182 183 184 185 186 187
PaddleOCR supports alternating training and evaluation. You can modify `eval_batch_step` in `configs/rec/rec_icdar15_train.yml` to set the evaluation frequency. By default, it is evaluated every 500 iter and the best acc model is saved under `output/rec_CRNN/best_accuracy` during the evaluation process.

If the evaluation set is large, the test will be time-consuming. It is recommended to reduce the number of evaluations, or evaluate after training.

* Tip: You can use the `-c` parameter to select multiple model configurations under the `configs/rec/` path for training. The recognition algorithms supported by PaddleOCR are:


| Configuration file |  Algorithm |   backbone |   trans   |   seq      |     pred     |
| :--------: |  :-------:   | :-------:  |   :-------:   |   :-----:   |  :-----:   |
188 189
| [rec_chinese_lite_train_v2.0.yml](../../configs/rec/ch_ppocr_v2.0/rec_chinese_lite_train_v2.0.yml) |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  |
| [rec_chinese_common_train_v2.0.yml](../../configs/rec/ch_ppocr_v2.0/rec_chinese_common_train_v2.0.yml) |  CRNN | ResNet34_vd |  None   |  BiLSTM |  ctc  |
K
Khanh Tran 已提交
190
| rec_chinese_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  |
W
WenmuZhou 已提交
191
| rec_chinese_common_train.yml |  CRNN |   ResNet34_vd |  None   |  BiLSTM |  ctc  |
K
Khanh Tran 已提交
192 193 194 195 196 197 198 199 200
| rec_icdar15_train.yml |  CRNN |   Mobilenet_v3 large 0.5 |  None   |  BiLSTM |  ctc  |
| rec_mv3_none_bilstm_ctc.yml |  CRNN |   Mobilenet_v3 large 0.5 |  None   |  BiLSTM |  ctc  |
| rec_mv3_none_none_ctc.yml |  Rosetta |   Mobilenet_v3 large 0.5 |  None   |  None |  ctc  |
| rec_mv3_tps_bilstm_ctc.yml |  STARNet |   Mobilenet_v3 large 0.5 |  tps   |  BiLSTM |  ctc  |
| rec_mv3_tps_bilstm_attn.yml |  RARE |   Mobilenet_v3 large 0.5 |  tps   |  BiLSTM |  attention  |
| rec_r34_vd_none_bilstm_ctc.yml |  CRNN |   Resnet34_vd |  None   |  BiLSTM |  ctc  |
| rec_r34_vd_none_none_ctc.yml |  Rosetta |   Resnet34_vd |  None   |  None |  ctc  |
| rec_r34_vd_tps_bilstm_ctc.yml | STARNet | Resnet34_vd | tps | BiLSTM | ctc |

W
WenmuZhou 已提交
201
For training Chinese data, it is recommended to use
202
[rec_chinese_lite_train_v2.0.yml](../../configs/rec/ch_ppocr_v2.0/rec_chinese_lite_train_v2.0.yml). If you want to try the result of other algorithms on the Chinese data set, please refer to the following instructions to modify the configuration file:
K
Khanh Tran 已提交
203
co
204
Take `rec_chinese_lite_train_v2.0.yml` as an example:
K
Khanh Tran 已提交
205 206 207
```
Global:
  ...
208 209
  # Add a custom dictionary, such as modify the dictionary, please point the path to the new dictionary
  character_dict_path: ppocr/utils/ppocr_keys_v1.txt
K
Khanh Tran 已提交
210 211 212
  # Modify character type
  character_type: ch
  ...
213
  # Whether to recognize spaces
214
  use_space_char: True
K
Khanh Tran 已提交
215

216 217 218 219

Optimizer:
  ...
  # Add learning rate decay strategy
220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264
  lr:
    name: Cosine
    learning_rate: 0.001
  ...

...

Train:
  dataset:
    # Type of dataset,we support LMDBDateSet and SimpleDataSet
    name: SimpleDataSet
    # Path of dataset
    data_dir: ./train_data/
    # Path of train list
    label_file_list: ["./train_data/train_list.txt"]
    transforms:
      ...
      - RecResizeImg:
          # Modify image_shape to fit long text
          image_shape: [3, 32, 320]
      ...
  loader:
    ...
    # Train batch_size for Single card
    batch_size_per_card: 256
    ...

Eval:
  dataset:
    # Type of dataset,we support LMDBDateSet and SimpleDataSet
    name: SimpleDataSet
    # Path of dataset
    data_dir: ./train_data
    # Path of eval list
    label_file_list: ["./train_data/val_list.txt"]
    transforms:
      ...
      - RecResizeImg:
          # Modify image_shape to fit long text
          image_shape: [3, 32, 320]
      ...
  loader:
    # Eval batch_size for Single card
    batch_size_per_card: 256
    ...
K
Khanh Tran 已提交
265 266 267
```
**Note that the configuration file for prediction/evaluation must be consistent with the training.**

W
WenmuZhou 已提交
268 269 270 271
<a name="Multi_language"></a>
- Multi-language

PaddleOCR also provides multi-language. The configuration file in `configs/rec/multi_languages` provides multi-language configuration files. Currently, the multi-language algorithms supported by PaddleOCR are:
K
Khanh Tran 已提交
272

W
WenmuZhou 已提交
273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289
| Configuration file | Algorithm name | backbone | trans | seq | pred | language |
| :--------: | :-------: | :-------: | :-------: | :-----: | :-----: | :-----: |
| rec_en_lite_train.yml | CRNN | Mobilenet_v3 small 0.5 | None | BiLSTM | ctc | English |
| rec_french_lite_train.yml | CRNN | Mobilenet_v3 small 0.5 | None | BiLSTM | ctc | French |
| rec_ger_lite_train.yml | CRNN | Mobilenet_v3 small 0.5 | None | BiLSTM | ctc | German |
| rec_japan_lite_train.yml | CRNN | Mobilenet_v3 small 0.5 | None | BiLSTM | ctc | Japanese |
| rec_korean_lite_train.yml | CRNN | Mobilenet_v3 small 0.5 | None | BiLSTM | ctc | Korean |

The multi-language model training method is the same as the Chinese model. The training data set is 100w synthetic data. A small amount of fonts and test data can be downloaded on [Baidu Netdisk](https://pan.baidu.com/s/1bS_u207Rm7YbY33wOECKDA),Extraction code:frgi.

If you want to finetune on the basis of the existing model effect, please refer to the following instructions to modify the configuration file:

Take `rec_french_lite_train` as an example:

```
Global:
  ...
290
  # Add a custom dictionary, such as modify the dictionary, please point the path to the new dictionary
W
WenmuZhou 已提交
291 292
  character_dict_path: ./ppocr/utils/dict/french_dict.txt
  ...
293
  # Whether to recognize spaces
294
  use_space_char: True
295

W
WenmuZhou 已提交
296
...
297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316

Train:
  dataset:
    # Type of dataset,we support LMDBDateSet and SimpleDataSet
    name: SimpleDataSet
    # Path of dataset
    data_dir: ./train_data/
    # Path of train list
    label_file_list: ["./train_data/french_train.txt"]
    ...

Eval:
  dataset:
    # Type of dataset,we support LMDBDateSet and SimpleDataSet
    name: SimpleDataSet
    # Path of dataset
    data_dir: ./train_data
    # Path of eval list
    label_file_list: ["./train_data/french_val.txt"]
    ...
W
WenmuZhou 已提交
317
```
K
Khanh Tran 已提交
318

W
WenmuZhou 已提交
319
<a name="EVALUATION"></a>
X
xxxpsyduck 已提交
320
### EVALUATION
K
Khanh Tran 已提交
321 322 323 324 325

The evaluation data set can be modified via `configs/rec/rec_icdar15_reader.yml` setting of `label_file_path` in EvalReader.

```
# GPU evaluation, Global.checkpoints is the weight to be tested
326
python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec_icdar15_reader.yml -o Global.checkpoints={path/to/weights}/best_accuracy
K
Khanh Tran 已提交
327 328
```

W
WenmuZhou 已提交
329
<a name="PREDICTION"></a>
X
xxxpsyduck 已提交
330
### PREDICTION
K
Khanh Tran 已提交
331

W
WenmuZhou 已提交
332
<a name="Training_engine_prediction"></a>
K
Khanh Tran 已提交
333 334 335 336 337 338 339 340
* Training engine prediction

Using the model trained by paddleocr, you can quickly get prediction through the following script.

The default prediction picture is stored in `infer_img`, and the weight is specified via `-o Global.checkpoints`:

```
# Predict English results
341
python3 tools/infer_rec.py -c configs/rec/ch_ppocr_v2.0/rec_chinese_lite_train_v2.0.yml -o Global.checkpoints={path/to/weights}/best_accuracy TestReader.infer_img=doc/imgs_words/en/word_1.jpg
K
Khanh Tran 已提交
342 343 344 345
```

Input image:

346
![](../imgs_words/en/word_1.png)
K
Khanh Tran 已提交
347 348 349 350 351 352 353 354 355

Get the prediction result of the input image:

```
infer_img: doc/imgs_words/en/word_1.png
     index: [19 24 18 23 29]
     word : joint
```

356
The configuration file used for prediction must be consistent with the training. For example, you completed the training of the Chinese model with `python3 tools/train.py -c configs/rec/ch_ppocr_v2.0/rec_chinese_lite_train_v2.0.yml`, you can use the following command to predict the Chinese model:
K
Khanh Tran 已提交
357 358 359

```
# Predict Chinese results
360
python3 tools/infer_rec.py -c configs/rec/ch_ppocr_v2.0/rec_chinese_lite_train_v2.0.yml -o Global.checkpoints={path/to/weights}/best_accuracy TestReader.infer_img=doc/imgs_words/ch/word_1.jpg
K
Khanh Tran 已提交
361 362 363 364
```

Input image:

365
![](../imgs_words/ch/word_1.jpg)
K
Khanh Tran 已提交
366 367 368 369 370 371 372 373

Get the prediction result of the input image:

```
infer_img: doc/imgs_words/ch/word_1.jpg
     index: [2092  177  312 2503]
     word : 韩国小馆
```