From 6735cced20bf905b88bb9fe9d63f811cca896481 Mon Sep 17 00:00:00 2001 From: zhiminzhang0830 <452516515@qq.com> Date: Thu, 29 Sep 2022 14:21:47 +0800 Subject: [PATCH] =?UTF-8?q?=E6=B7=BB=E5=8A=A0RFL=E6=96=87=E6=A1=A3?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- doc/doc_ch/algorithm_overview.md | 2 + doc/doc_ch/algorithm_rec_rfl.md | 161 ++++++++++++++++++++++++++++ doc/doc_en/algorithm_overview_en.md | 3 +- doc/doc_en/algorithm_rec_rfl_en.md | 143 ++++++++++++++++++++++++ 4 files changed, 308 insertions(+), 1 deletion(-) create mode 100644 doc/doc_ch/algorithm_rec_rfl.md create mode 100644 doc/doc_en/algorithm_rec_rfl_en.md diff --git a/doc/doc_ch/algorithm_overview.md b/doc/doc_ch/algorithm_overview.md index ecb0e9df..4351fdbf 100755 --- a/doc/doc_ch/algorithm_overview.md +++ b/doc/doc_ch/algorithm_overview.md @@ -79,6 +79,7 @@ PaddleOCR将**持续新增**支持OCR领域前沿算法与模型，**欢迎广 - [x] [VisionLAN](./algorithm_rec_visionlan.md) - [x] [SPIN](./algorithm_rec_spin.md) - [x] [RobustScanner](./algorithm_rec_robustscanner.md) +- [x] [RFL](./algorithm_rec_rfl.md) 参考[DTRB](https://arxiv.org/abs/1904.01906)[3]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： @@ -102,6 +103,7 @@ PaddleOCR将**持续新增**支持OCR领域前沿算法与模型，**欢迎广 |VisionLAN|Resnet45| 90.30% | rec_r45_visionlan | [训练模型](https://paddleocr.bj.bcebos.com/rec_r45_visionlan_train.tar) | |SPIN|ResNet32| 90.00% | rec_r32_gaspin_bilstm_att | coming soon | |RobustScanner|ResNet31| 87.77% | rec_r31_robustscanner | coming soon | +|RFL|ResNetRFL| 88.63% | rec_resnet_rfl_att | [训练模型](https://paddleocr.bj.bcebos.com/contribution/rec_resnet_rfl.tar) | diff --git a/doc/doc_ch/algorithm_rec_rfl.md b/doc/doc_ch/algorithm_rec_rfl.md new file mode 100644 index 00000000..5135e77a --- /dev/null +++ b/doc/doc_ch/algorithm_rec_rfl.md @@ -0,0 +1,161 @@ +# 场景文本识别算法-RFL + +- [1. 算法简介](#1) +- [2. 环境配置](#2) +- [3. 模型训练、评估、预测](#3) + - [3.1 训练](#3-1) + - [3.2 评估](#3-2) + - [3.3 预测](#3-3) +- [4. 推理部署](#4) + - [4.1 Python推理](#4-1) + - [4.2 C++推理](#4-2) + - [4.3 Serving服务化部署](#4-3) + - [4.4 更多推理部署](#4-4) +- [5. FAQ](#5) + + +## 1. 算法简介 + +论文信息： +> [Reciprocal Feature Learning via Explicit and Implicit Tasks in Scene Text Recognition](https://arxiv.org/abs/2105.06229.pdf) +> Hui Jiang, Yunlu Xu, Zhanzhan Cheng, Shiliang Pu, Yi Niu, Wenqi Ren, Fei Wu, and Wenming Tan +> ICDAR, 2021 + + + +`RFL`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，算法复现效果如下： + +|模型|骨干网络|配置文件|Acc|下载链接| +| --- | --- | --- | --- | --- | +|RFL-CNT|ResNetRFL|[rec_resnet_rfl_visual.yml](../../configs/rec/rec_resnet_rfl_visual.yml)|93.40%|[训练模型](https://paddleocr.bj.bcebos.com/contribution/rec_resnet_rfl.tar)| +|RFL-Att|ResNetRFL|[rec_resnet_rfl_att.yml](../../configs/rec/rec_resnet_rfl_att.yml)|88.63%|[训练模型](https://paddleocr.bj.bcebos.com/contribution/rec_resnet_rfl.tar)| + + +## 2. 环境配置 +请先参考[《运行环境准备》](./environment.md)配置PaddleOCR运行环境，参考[《项目克隆》](./clone.md)克隆项目代码。 + + + +## 3. 模型训练、评估、预测 + + +### 3.1 模型训练 + +请参考[文本识别训练教程](./recognition.md)。PaddleOCR对代码进行了模块化，训练`RFL`识别模型时需要**更换配置文件**为`RFL`的[配置文件](../../configs/rec/rec_resnet_rfl_att.yml)。 + +#### 启动训练 + + +具体地，在完成数据准备后，便可以启动训练，训练命令如下： +```shell +#step1:训练CNT分支 +#单卡训练（训练周期长，不建议） +python3 tools/train.py -c configs/rec/rec_resnet_rfl_visual.yml + +#多卡训练，通过--gpus参数指定卡号 +python3 -m paddle.distributed.launch --gpus '0,1,2,3' tools/train.py -c configs/rec/rec_resnet_rfl_visual.yml + +#step2:联合训练CNT和Att分支,注意将pretrained_model的路径设置为本地路径。 +#单卡训练（训练周期长，不建议） +python3 tools/train.py -c configs/rec/rec_resnet_rfl_att.yml -o Global.pretrained_model=./output/rec/rec_resnet_rfl_visual/best_accuracy + +#多卡训练，通过--gpus参数指定卡号 +python3 -m paddle.distributed.launch --gpus '0,1,2,3' tools/train.py -c configs/rec/rec_resnet_rfl_att.yml -o Global.pretrained_model=./output/rec/rec_resnet_rfl_visual/best_accuracy +``` + + +### 3.2 评估 + +可下载已训练完成的[模型文件](https://paddleocr.bj.bcebos.com/contribution/rec_resnet_rfl.tar)，使用如下命令进行评估： + +```shell +# 注意将pretrained_model的路径设置为本地路径。 +python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec_resnet_rfl_att.yml -o Global.pretrained_model=./output/rec/rec_resnet_rfl_att/best_accuracy +``` + + +### 3.3 预测 + +使用如下命令进行单张图片预测： +```shell +# 注意将pretrained_model的路径设置为本地路径。 +python3 tools/infer_rec.py -c configs/rec/rec_resnet_rfl_att.yml -o Global.infer_img='./doc/imgs_words_en/word_10.png' Global.pretrained_model=./output/rec/rec_resnet_rfl_att/best_accuracy +# 预测文件夹下所有图像时，可修改infer_img为文件夹，如 Global.infer_img='./doc/imgs_words_en/'。 +``` + + + +## 4. 推理部署 + + +### 4.1 Python推理 +首先将训练得到best模型，转换成inference model。这里以训练完成的模型为例（[模型下载地址](https://paddleocr.bj.bcebos.com/contribution/rec_resnet_rfl.tar) )，可以使用如下命令进行转换： + +```shell +# 注意将pretrained_model的路径设置为本地路径。 +python3 tools/export_model.py -c configs/rec/rec_resnet_rfl_att.yml -o Global.pretrained_model=./output/rec/rec_resnet_rfl_att/best_accuracy Global.save_inference_dir=./inference/rec_resnet_rfl_att/ +``` +**注意：** +- 如果您是在自己的数据集上训练的模型，并且调整了字典文件，请注意修改配置文件中的`character_dict_path`是否是所需要的字典文件。 +- 如果您修改了训练时的输入大小，请修改`tools/export_model.py`文件中的对应RFL的`infer_shape`。 + +转换成功后，在目录下有三个文件： +``` +/inference/rec_resnet_rfl_att/ + ├── inference.pdiparams # 识别inference模型的参数文件 + ├── inference.pdiparams.info # 识别inference模型的参数信息，可忽略 + └── inference.pdmodel # 识别inference模型的program文件 +``` + +执行如下命令进行模型推理： + +```shell +python3 tools/infer/predict_rec.py --image_dir='./doc/imgs_words_en/word_10.png' --rec_model_dir='./inference/rec_resnet_rfl_att/' --rec_algorithm='RFL' --rec_image_shape='1,32,100' +# 预测文件夹下所有图像时，可修改image_dir为文件夹，如 --image_dir='./doc/imgs_words_en/'。 +``` + +![](../imgs_words_en/word_10.png) + +执行命令后，上面图像的预测结果（识别的文本和得分）会打印到屏幕上，示例如下： +结果如下： +```shell +Predicts of ./doc/imgs_words_en/word_10.png:('pain', 0.9999927282333374) +``` + +**注意**： + +- 训练上述模型采用的图像分辨率是[1，32，100]，需要通过参数`rec_image_shape`设置为您训练时的识别图像形状。 +- 在推理时需要设置参数`rec_char_dict_path`指定字典，如果您修改了字典，请修改该参数为您的字典文件。 +- 如果您修改了预处理方法，需修改`tools/infer/predict_rec.py`中RFL的预处理为您的预处理方法。 + + + +### 4.2 C++推理部署 + +由于C++预处理后处理还未支持RFL，所以暂未支持 + + +### 4.3 Serving服务化部署 + +暂不支持 + + +### 4.4 更多推理部署 + +暂不支持 + + +## 5. FAQ + + +## 引用 + +```bibtex +@article{2021Reciprocal, + title = {Reciprocal Feature Learning via Explicit and Implicit Tasks in Scene Text Recognition}, + author = {Jiang, H. and Xu, Y. and Cheng, Z. and Pu, S. and Niu, Y. and Ren, W. and Wu, F. and Tan, W. }, + booktitle = {ICDAR}, + year = {2021}, + url = {https://arxiv.org/abs/2105.06229} +} +``` diff --git a/doc/doc_en/algorithm_overview_en.md b/doc/doc_en/algorithm_overview_en.md index bca22f78..f7ef7ad4 100755 --- a/doc/doc_en/algorithm_overview_en.md +++ b/doc/doc_en/algorithm_overview_en.md @@ -76,6 +76,7 @@ Supported text recognition algorithms (Click the link to get the tutorial): - [x] [VisionLAN](./algorithm_rec_visionlan_en.md) - [x] [SPIN](./algorithm_rec_spin_en.md) - [x] [RobustScanner](./algorithm_rec_robustscanner_en.md) +- [x] [RFL](./algorithm_rec_rfl_en.md) 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: @@ -99,7 +100,7 @@ Refer to [DTRB](https://arxiv.org/abs/1904.01906), the training and evaluation r |VisionLAN|Resnet45| 90.30% | rec_r45_visionlan | [trained model](https://paddleocr.bj.bcebos.com/rec_r45_visionlan_train.tar) | |SPIN|ResNet32| 90.00% | rec_r32_gaspin_bilstm_att | coming soon | |RobustScanner|ResNet31| 87.77% | rec_r31_robustscanner | coming soon | - +|RFL|ResNetRFL| 88.63% | rec_resnet_rfl_att | [trained model](https://paddleocr.bj.bcebos.com/contribution/rec_resnet_rfl.tar) | diff --git a/doc/doc_en/algorithm_rec_rfl_en.md b/doc/doc_en/algorithm_rec_rfl_en.md new file mode 100644 index 00000000..8f0adfbe --- /dev/null +++ b/doc/doc_en/algorithm_rec_rfl_en.md @@ -0,0 +1,143 @@ +# RFL + +- [1. Introduction](#1) +- [2. Environment](#2) +- [3. Model Training / Evaluation / Prediction](#3) + - [3.1 Training](#3-1) + - [3.2 Evaluation](#3-2) + - [3.3 Prediction](#3-3) +- [4. Inference and Deployment](#4) + - [4.1 Python Inference](#4-1) + - [4.2 C++ Inference](#4-2) + - [4.3 Serving](#4-3) + - [4.4 More](#4-4) +- [5. FAQ](#5) + + +## 1. Introduction + +Paper: +> [Reciprocal Feature Learning via Explicit and Implicit Tasks in Scene Text Recognition](https://arxiv.org/abs/2105.06229.pdf) +> Hui Jiang, Yunlu Xu, Zhanzhan Cheng, Shiliang Pu, Yi Niu, Wenqi Ren, Fei Wu, and Wenming Tan +> ICDAR, 2021 + +Using MJSynth and SynthText two text recognition datasets for training, and evaluating on IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE datasets, the algorithm reproduction effect is as follows: + +|Model|Backbone|config|Acc|Download link| +| --- | --- | --- | --- | --- | +|RFL-CNT|ResNetRFL|[rec_resnet_rfl_visual.yml](../../configs/rec/rec_resnet_rfl_visual.yml)|93.40%|[训练模型](https://paddleocr.bj.bcebos.com/contribution/rec_resnet_rfl.tar)| +|RFL-Att|ResNetRFL|[rec_resnet_rfl_att.yml](../../configs/rec/rec_resnet_rfl_att.yml)|88.63%|[训练模型](https://paddleocr.bj.bcebos.com/contribution/rec_resnet_rfl.tar)| + + +## 2. Environment +Please refer to ["Environment Preparation"](./environment_en.md) to configure the PaddleOCR environment, and refer to ["Project Clone"](./clone_en.md) to clone the project code. + + + +## 3. Model Training / Evaluation / Prediction + +Please refer to [Text Recognition Tutorial](./recognition_en.md). PaddleOCR modularizes the code, and training different recognition models only requires **changing the configuration file**. + +Training: + +Specifically, after the data preparation is completed, the training can be started. The training command is as follows: + +``` +#step1:train the CNT branch +#Single GPU training (long training period, not recommended) +python3 tools/train.py -c configs/rec/rec_resnet_rfl_visual.yml + +#Multi GPU training, specify the gpu number through the --gpus parameter +python3 -m paddle.distributed.launch --gpus '0,1,2,3' tools/train.py -c configs/rec/rec_resnet_rfl_visual.yml + +#step2:joint training of CNT and Att branches +#Single GPU training (long training period, not recommended) +python3 tools/train.py -c configs/rec/rec_resnet_rfl_att.yml -o Global.pretrained_model={path/to/weights}/best_accuracy + +#Multi GPU training, specify the gpu number through the --gpus parameter +python3 -m paddle.distributed.launch --gpus '0,1,2,3' tools/train.py -c configs/rec/rec_resnet_rfl_att.yml -o Global.pretrained_model={path/to/weights}/best_accuracy + + +``` + +Evaluation: + +``` +# GPU evaluation +python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec_resnet_rfl_att.yml -o Global.pretrained_model={path/to/weights}/best_accuracy +``` + +Prediction: + +``` +# The configuration file used for prediction must match the training +python3 tools/infer_rec.py -c configs/rec/rec_resnet_rfl_att.yml -o Global.infer_img='./doc/imgs_words_en/word_10.png' Global.pretrained_model={path/to/weights}/best_accuracy +``` + + +## 4. Inference and Deployment + + +### 4.1 Python Inference +First, the model saved during the RFL text recognition training process is converted into an inference model. ( [Model download link](https://paddleocr.bj.bcebos.com/contribution/rec_resnet_rfl.tar)) ), you can use the following command to convert: + +``` +python3 tools/export_model.py -c configs/rec/rec_resnet_rfl_att.yml -o Global.pretrained_model={path/to/weights}/best_accuracy Global.save_inference_dir=./inference/rec_resnet_rfl_att +``` + +**Note:** +- If you are training the model on your own dataset and have modified the dictionary file, please pay attention to modify the `character_dict_path` in the configuration file to the modified dictionary file. +- If you modified the input size during training, please modify the `infer_shape` corresponding to NRTR in the `tools/export_model.py` file. + +After the conversion is successful, there are three files in the directory: +``` +/inference/rec_resnet_rfl_att/ + ├── inference.pdiparams + ├── inference.pdiparams.info + └── inference.pdmodel +``` + + +For RFL text recognition model inference, the following commands can be executed: + +``` +python3 tools/infer/predict_rec.py --image_dir='./doc/imgs_words_en/word_10.png' --rec_model_dir='./inference/rec_resnet_rfl_att/' --rec_algorithm='RFL' --rec_image_shape='1,32,100' +``` + +![](../imgs_words_en/word_10.png) + +After executing the command, the prediction result (recognized text and score) of the image above is printed to the screen, an example is as follows: +The result is as follows: +```shell +Predicts of ./doc/imgs_words_en/word_10.png:('pain', 0.9999927282333374) +``` + + +### 4.2 C++ Inference + +Not supported + + +### 4.3 Serving + +Not supported + + +### 4.4 More + +Not supported + + +## 5. FAQ + +## Citation + +```bibtex +@article{2021Reciprocal, + title = {Reciprocal Feature Learning via Explicit and Implicit Tasks in Scene Text Recognition}, + author = {Jiang, H. and Xu, Y. and Cheng, Z. and Pu, S. and Niu, Y. and Ren, W. and Wu, F. and Tan, W. }, + booktitle = {ICDAR}, + year = {2021}, + url = {https://arxiv.org/abs/2105.06229} +} +``` -- GitLab