README.md · haohao_wh/mindocr

# MindOCR

[![CI](https://github.com/mindspore-lab/mindocr/actions/workflows/ci.yml/badge.svg)](https://github.com/mindspore-lab/mindocr/actions/workflows/ci.yml)
[![license](https://img.shields.io/github/license/mindspore-lab/mindocr.svg)](https://github.com/mindspore-lab/mindocr/blob/main/LICENSE)
[![open issues](https://img.shields.io/github/issues/mindspore-lab/mindocr)](https://github.com/mindspore-lab/mindocr/issues)
[![PRs](https://img.shields.io/badge/PRs-welcome-pink.svg)](https://github.com/mindspore-lab/mindocr/pulls)
[![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)

English | [中文](README_CN.md)

</div>

## Introduction
MindOCR is an open-source toolbox for OCR development and application based on [MindSpore](https://www.mindspore.cn/en), which integrates series of mainstream text detection and recognition algorihtms and models and provides easy-to-use training and inference tools. It can accelerate the process of developing and deploying SoTA text detection and recognition models in real-world applications, such as DBNet/DBNet++ and CRNN/SVTR, and help fulfill the need of image-text understanding .

<details open>
<summary> Major Features </summary>

- **Modular design**: We decoupled the OCR task into several configurable modules. Users can setup the training and evaluation pipelines, customize the data processing pipeline and model architectures easily by modifying just few lines of code.
- **High-performance**: MindOCR provides a series of pretrained weights trained with optimized configurations that reach competitive performance on OCR tasks.
- **Low-cost-to-apply**: Easy-to-use inference tools are provided in MindOCR to perform text detection and recognition tasks.
</details>

## Installation

#### Prerequisites

MindOCR is built on MindSpore AI framework, which supports CPU/GPU/NPU devices.
MindOCR is compatible with the following framework versions. For details and installation guideline, please refer to the installation links shown below.

- mindspore >= 1.9  [[install](https://www.mindspore.cn/install)]
- python >= 3.7
- openmpi 4.0.3 (for distributed training/evaluation)  [[install](https://www.open-mpi.org/software/ompi/v4.0/)]
- mindspore lite (for inference)  [[install](docs/en/inference/environment_en.md)]

#### Dependency
```shell
pip install -r requirements.txt
```
**Tips:**

- If scikit_image cannot be imported, please set environment variable `$LD_PRELOAD`
as follows, (related [opencv issue](https://github.com/opencv/opencv/issues/14884))

```shell
    export LD_PRELOAD=path/to/scikit_image.libs/libgomp-d22c30c5.so.1.0.0:$LD_PRELOAD
    ```

#### Install from Source (recommend)
```shell
git clone https://github.com/mindspore-lab/mindocr.git
cd mindocr
pip install -e .
```
> Using `-e` for "editable" mode can help resolve potential module import issues.

#### Install from PyPI
```shell
pip install mindocr
```
> As this project is under active development, the version installed from PyPI is out-of-date currently. (will update soon).

## Quick Start

### Text Detection and Recognition Demo

After installing MindOCR, we can run text detection and recognition on an arbitrary image easily as follows.

```shell
python tools/infer/text/predict_system.py --image_dir {path_to_img or dir_to_imgs} \
                                          --det_algorithm DB++  \
                                          --rec_algorithm CRNN
```

After running, the results will be saved in `./inference_results` by default. Here is an example result.

<p align="center">
  <img src="https://github.com/SamitHuang/mindocr-1/assets/8156835/c1f53970-8618-4039-994f-9f6dc1eee1dd" width=600 />
</p>
<p align="center">
  <em> Visualization of text detection and recognition result </em>
</p>

We can see that all texts on the image are detected and recognized accurately. For more usage, please refer to the inference section in [tutorials](#tutorials).

### Model Training and Evaluation - Quick Guideline

It is easy to train your OCR model with the `tools/train.py` script, which supports both text detection and recognition model training.

```shell
python tools/train.py --config {path/to/model_config.yaml}
```

The `--config` arg specifies the path to a yaml file that defines the model to be trained and the training strategy including data process pipeline, optimizer, lr scheduler, etc.

MindOCR provides SoTA OCR models with their training strategies in `configs` folder.
You may adapt it to your task/dataset, for example, by running

```shell
# train text detection model DBNet++ on icdar15 dataset
python tools/train.py --config configs/det/dbnet/db++_r50_icdar15.yaml
```
```shell
# train text recognition model CRNN on icdar15 dataset
python tools/train.py --config configs/rec/crnn/crnn_icdar15.yaml
```

Similarly, it is simple to evaluate the trained model with the `tools/eval.py` script.
```shell
python tools/eval.py \
    --config {path/to/model_config.yaml} \
    --opt eval.dataset_root={path/to/your_dataset} eval.ckpt_load_path={path/to/ckpt_file}
```

For more illustration and usage, please refer to the model training section in [Tutorials](#tutorials).

## Tutorials

- Datasets
    - [Dataset Preparation](tools/dataset_converters/README.md)
    - [Data Transformation Mechanism](docs/en/tutorials/transform_tutorial.md)
- Model Training
    - [Yaml Configuration](docs/en/tutorials/yaml_configuration.md)
    - [Text Detection]()  (coming soon)
    - [Text Recognition](docs/en/tutorials/training_recognition_custom_dataset.md)
    - [Distributed Training](docs/cn/tutorials/distribute_train.md)
    - [Advance: Gradient Accumulation, EMA, Resume Training, etc](docs/en/tutorials/advanced_train.md)
- Inference and Deployment
    - [Python/C++ Inference on Ascend 310](docs/en/inference/inference_tutorial_en.md)
    - [Python Online Inference](tools/infer/text/README.md)
- Developer Guides
    - [Customize Dataset](mindocr/data/README.md)
    - [Customize Data Transformation](mindocr/data/transforms/README.md)
    - [Customize a New Model](mindocr/models/README.md)
    - [Customize Postprocessing Method](mindocr/postprocess/README.md)

## Model List

<details open>
<summary>Text Detection</summary>

- [x] [DBNet](configs/det/dbnet/README.md) (AAAI'2020)
- [x] [DBNet++](configs/det/dbnet/README.md) (TPAMI'2022)
- [x] [PSENet](configs/det/psenet/README.md) (CVPR'2019)
- [x] [EAST](configs/det/east/README.md)(CVPR'2017)
- [ ] [FCENet](https://arxiv.org/abs/2104.10442) (CVPR'2021) [coming soon]

</details>

<details open>
<summary>Text Recognition</summary>

- [x] [CRNN](configs/rec/crnn/README.md) (TPAMI'2016)
- [x] [CRNN-Seq2Seq/RARE](configs/rec/rare/README.md) (CVPR'2016)
- [x] [SVTR](configs/rec/svtr/README.md) (IJCAI'2022)
- [ ] [ABINet](https://arxiv.org/abs/2103.06495) (CVPR'2021) [coming soon]

</details>

For the detailed performance of the trained models, please refer to [configs](./configs).

For detailed support for MindSpore Lite and ACL inference models, please refer to [MindOCR Models Support List](docs/en/inference/models_list_en.md) and [Third-Party Models Support List](docs/en/inference/models_list_thirdparty_en.md).

## Dataset List

MindOCR provides a [dataset conversion tool](tools/dataset_converters) to OCR datasets with different formats and support customized dataset by users. We have validated the following public OCR datasets in model training/evaluation.

<details open>
<summary>General OCR Datasets</summary>

- [x] [ICDAR2015](https://rrc.cvc.uab.es/?ch=4) [[paper](https://rrc.cvc.uab.es/files/short_rrc_2015.pdf)] [[download](docs/en/datasets/icdar2015.md)]

- [x] [Total-Text](https://github.com/cs-chan/Total-Text-Dataset/tree/master/Dataset)  [[paper](https://arxiv.org/abs/1710.10400)]  [[download](docs/en/datasets/totaltext.md)]

- [x] [Syntext150k](https://github.com/aim-uofa/AdelaiDet) [[paper](https://arxiv.org/abs/2002.10200)] [[download](docs/en/datasets/syntext150k.md)]

- [x] [MLT2017](https://rrc.cvc.uab.es/?ch=8&com=introduction) [[paper](https://ieeexplore.ieee.org/abstract/document/8270168)]  [[download](docs/en/datasets/mlt2017.md)] (multi-language)

- [x] [MSRA-TD500](http://www.iapr-tc11.org/mediawiki/index.php/MSRA_Text_Detection_500_Database_(MSRA-TD500)) [[paper](https://ieeexplore.ieee.org/abstract/document/6247787)]  [[download](docs/en/datasets/td500.md)]

- [x] [SCUT-CTW1500](https://github.com/Yuliang-Liu/Curve-Text-Detector) [[paper](https://www.sciencedirect.com/science/article/pii/S0031320319300664)]   [[download](docs/en/datasets/ctw1500.md)]

- [x] [Chinese-Text-Recognition-Benchmark](https://github.com/FudanVI/benchmarking-chinese-text-recognition)  [[paper](https://arxiv.org/abs/2112.15093)]   [[download](https://github.com/FudanVI/benchmarking-chinese-text-recognition#download)]

</details>

We will include more datasets for training and evaluation. This list will be continuously updated.

## Notes

### What is New

- 2023/06/07
1. Add new trained models
    - [PSENet](configs/det/psenet) for text detection
    - [EAST](configs/det/east) for text detection
    - [SVTR](configs/rec/svtr) for text recognition
2. Add more benchmark datasets and their results
    - [totaltext](docs/en/datasets/totaltext.md)
    - [mlt2017](docs/en/datasets/mlt2017.md)
    - [chinese_text_recognition](docs/en/datasets/chinese_text_recognition.md)
3. Add resume training function, which can be used in case of unexpected interruption in training. Usage: add the `resume` parameter under the `model` field in the yaml config, e.g.,`resume: True`, load and resume training from {ckpt_save_dir}/train_resume.ckpt or `resume: /path/to/train_resume.ckpt`, load and resume training from the given path.
4. Improve postprocessing for detection: re-scale detected text polygons to original image space by default,
which can be enabled by add "shape_list" to the `eval.dataset.output_columns` list.
5. Refactor online inference to support more models, see [README.md](tools/infer/text/README.md) for details.

- 2023/05/15
1. Add new trained models
    - [DBNet++](configs/det/dbnet) for text detection
    - [CRNN-Seq2Seq](configs/rec/rare) for text recognition
    - DBNet pretrained on SynthText is now available: [checkpoint url](https://download.mindspore.cn/toolkits/mindocr/dbnet/dbnet_resnet50_synthtext-40655acb.ckpt)
2. Add more benchmark datasets and their results
    - [SynthText](https://academictorrents.com/details/2dba9518166cbd141534cbf381aa3e99a087e83c), [MSRA-TD500](docs/en/datasets/td500.md), [CTW1500](docs/en/datasets/ctw1500.md)
    - More benchmark results for DBNet are reported [here](configs/det/dbnet/README.md).
3. Add checkpoint manager for saving top-k checkpoints and improve log.
4. Python inference code refactored.
5. Bug fix: use Meter to average loss for large datasets, disable `pred_cast_fp32` for ctcloss in AMP training, fix error when invalid polygons exist.

- 2023/05/04
1. Support loading self-defined pretrained checkpoints via setting `model-pretrained` with checkpoint url or local path in yaml.
2. Support setting probability for executing augmentation including rotation and flip.
3. Add Exponential Moving Average(EMA) for model training, which can be enabled by setting `train-ema` (default: False) and `train-ema_decay` in the yaml config.
4. Arg parameter changed：`num_columns_to_net` -> `net_input_column_index`: change the column number feeding into the network to the column index.
5. Arg parameter changed：`num_columns_of_labels` -> `label_column_index`: change the column number corresponds to the label to the column index.

- 2023/04/21
1. Add parameter grouping to support flexible regularization in training. Usage: add `grouping_strategy` argument in yaml config to select a predefined grouping strategy, or use `no_weight_decay_params` argument to pick layers to exclude from weight decay (e.g., bias, norm). Example can be referred in `configs/rec/crnn/crnn_icdar15.yaml`
2. Add gradient accumulation to support large batch size training. Usage: add `gradient_accumulation_steps` in yaml config, the global batch size = batch_size * devices * gradient_accumulation_steps. Example can be referred in `configs/rec/crnn/crnn_icdar15.yaml`
3. Add gradient clip to support training stablization. Enable it by setting `grad_clip` as True in yaml config.

- 2023/03/23
1. Add dynamic loss scaler support, compatible with drop overflow update. To enable dynamic loss scaler, please set `type` of `loss_scale` as `dynamic`. A YAML example can be viewed in `configs/rec/crnn/crnn_icdar15.yaml`

- 2023/03/20
1. Arg names changed: `output_keys` -> `output_columns`, `num_keys_to_net` -> `num_columns_to_net`
2. Data pipeline updated

- 2023/03/13
1. Add system test and CI workflow.
2. Add modelarts adapter to allow training on OpenI platform. To train on OpenI:
  ```text
    i)   Create a new training task on the openi cloud platform.
    ii)  Link the dataset (e.g., ic15_mindocr) on the webpage.
    iii) Add run parameter `config` and write the yaml file path on the website UI interface, e.g., '/home/work/user-job-dir/V0001/configs/rec/test.yaml'
    iv)  Add run parameter `enable_modelarts` and set True on the website UI interface.
    v)   Fill in other blanks and launch.
  ```

### How to Contribute

We appreciate all kinds of contributions including issues and PRs to make MindOCR better.

Please refer to [CONTRIBUTING.md](CONTRIBUTING.md) for the contributing guideline. Please follow the [Model Template and Guideline](mindocr/models/README.md) for contributing a model that fits the overall interface :)

### License

This project follows the [Apache License 2.0](LICENSE) open-source license.

### Citation

If you find this project useful in your research, please consider citing:

```latex
@misc{MindSpore OCR 2023,
    title={{MindSpore OCR }:MindSpore OCR Toolbox},
    author={MindSpore Team},
    howpublished = {\url{https://github.com/mindspore-lab/mindocr/}},
    year={2023}
}
```

haohao_wh/mindocr

简介

发行版

贡献者

近期动态

haohao_wh/mindocr .gitee-modal { width: 500px !important; }

简介

发行版

贡献者

近期动态

搜索帮助

haohao_wh/mindocr