提交 7b928353 编写于 作者: T tink2123

format angle_class

上级 a671fdfb
## 文字角度分类
### 方法介绍
文字角度分类主要用于图片非0度的场景下,在这种场景下需要对图片里检测到的文本行进行一个转正的操作。在PaddleOCR系统内,
# 文本方向分类器
- [方法介绍](#方法介绍)
- [数据准备](#数据准备)
- [启动训练](#启动训练)
- [训练](#训练)
- [评估](#评估)
- [预测](#预测)
<a name="方法介绍"></a>
## 方法介绍
文本方向分类器主要用于图片非0度的场景下,在这种场景下需要对图片里检测到的文本行进行一个转正的操作。在PaddleOCR系统内,
文字检测之后得到的文本行图片经过仿射变换之后送入识别模型,此时只需要对文字进行一个0和180度的角度分类,因此PaddleOCR内置的
字角度分类器**只支持了0和180度的分类**。如果想支持更多角度,可以自己修改算法进行支持。
本方向分类器**只支持了0和180度的分类**。如果想支持更多角度,可以自己修改算法进行支持。
0和180度数据样本例子:
![](../imgs_results/angle_class_example.jpg)
### 数据准备
<a name="数据准备"></a>
## 数据准备
请按如下步骤设置数据集:
......@@ -59,8 +69,8 @@ train/cls/train/word_002.jpg 180
|- word_003.jpg
| ...
```
### 启动训练
<a name="启动训练"></a>
## 启动训练
将准备好的txt文件和图片文件夹路径分别写入配置文件的 `Train/Eval.dataset.label_file_list``Train/Eval.dataset.data_dir` 字段下,`Train/Eval.dataset.data_dir`字段下的路径和文件里记载的图片名构成了图片的绝对路径。
......@@ -88,7 +98,8 @@ PaddleOCR提供了多种数据增强方式,如果您希望在训练时加入
*由于OpenCV的兼容性问题,扰动操作暂时只支持linux*
### 训练
<a name="训练"></a>
## 训练
PaddleOCR支持训练和评估交替进行, 可以在 `configs/cls/cls_mv3.yml` 中修改 `eval_batch_step` 设置评估频率,默认每1000个iter评估一次。训练过程中将会保存如下内容:
```bash
......@@ -106,7 +117,8 @@ PaddleOCR支持训练和评估交替进行, 可以在 `configs/cls/cls_mv3.yml`
**注意,预测/评估时的配置文件请务必与训练一致。**
### 评估
<a name="评估"></a>
## 评估
评估数据集可以通过修改`configs/cls/cls_mv3.yml`文件里的`Eval.dataset.label_file_list` 字段设置。
......@@ -116,7 +128,8 @@ export CUDA_VISIBLE_DEVICES=0
python3 tools/eval.py -c configs/cls/cls_mv3.yml -o Global.checkpoints={path/to/weights}/best_accuracy
```
### 预测
<a name="预测"></a>
## 预测
* 训练引擎的预测
......
## TEXT ANGLE CLASSIFICATION
# TEXT ANGLE CLASSIFICATION
### Method introduction
- [Method introduction](#method-introduction)
- [DATA PREPARATION](#data-preparation)
- [TRAINING](#training)
- [EVALUATION](#evaluation)
- [PREDICTION](#prediction)
<a name="method-introduction"></a>
## Method introduction
The angle classification is used in the scene where the image is not 0 degrees. In this scene, it is necessary to perform a correction operation on the text line detected in the picture. In the PaddleOCR system,
The text line image obtained after text detection is sent to the recognition model after affine transformation. At this time, only a 0 and 180 degree angle classification of the text is required, so the built-in PaddleOCR text angle classifier **only supports 0 and 180 degree classification**. If you want to support more angles, you can modify the algorithm yourself to support.
Example of 0 and 180 degree data samples:
![](../imgs_results/angle_class_example.jpg)
### DATA PREPARATION
<a name="data-preparation"></a>
## DATA PREPARATION
Please organize the dataset as follows:
......@@ -62,8 +71,8 @@ containing all images (test) and a cls_gt_test.txt. The structure of the test se
|- word_003.jpg
| ...
```
### TRAINING
<a name="training"></a>
## TRAINING
Write the prepared txt file and image folder path into the configuration file under the `Train/Eval.dataset.label_file_list` and `Train/Eval.dataset.data_dir` fields, the absolute path of the image consists of the `Train/Eval.dataset.data_dir` field and the image name recorded in the txt file.
PaddleOCR provides training scripts, evaluation scripts, and prediction scripts.
......@@ -107,7 +116,8 @@ If the evaluation set is large, the test will be time-consuming. It is recommend
**Note that the configuration file for prediction/evaluation must be consistent with the training.**
### EVALUATION
<a name="evaluation"></a>
## EVALUATION
The evaluation dataset can be set by modifying the `Eval.dataset.label_file_list` field in the `configs/cls/cls_mv3.yml` file.
......@@ -116,8 +126,8 @@ export CUDA_VISIBLE_DEVICES=0
# GPU evaluation, Global.checkpoints is the weight to be tested
python3 tools/eval.py -c configs/cls/cls_mv3.yml -o Global.checkpoints={path/to/weights}/best_accuracy
```
### PREDICTION
<a name="prediction"></a>
## PREDICTION
* Training engine prediction
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册