EasyOCR

https://www.jaided.ai/easyocr/
Admin
EasyOCR is a Python module for extracting text from images, providing optical character recognition...
Tokens:5,175
Snippets:37
Trust Score:8.7
Update:4 weeks ago
Show doc for...
Tokens:
Raw
### Install EasyOCR from source

Source: https://www.jaided.ai/easyocr/install

Installs the latest development version directly from the GitHub repository.

```bash
pip install git+https://github.com/JaidedAI/EasyOCR.git
```

--------------------------------

### Install EasyOCR via pip

Source: https://www.jaided.ai/easyocr/install

Installs the latest stable EasyOCR release from PyPI. Recommended for most users.

```bash
pip install easyocr
```

--------------------------------

### Get simple text output with detail=0

Source: https://www.jaided.ai/easyocr/tutorial

The detail=0 parameter turns off the default detailed output, returning only a list of recognized text strings.

```python
reader.readtext('chinese_tra.jpg', detail=0)
# ['高鐵左營站', 'HSR', 'Station', '汽車臨停接送區', 'Kiss', 'Car', 'and', 'Ride']
```

--------------------------------

### Import easyocr

Source: https://www.jaided.ai/easyocr/tutorial

Imports the EasyOCR library.

```python
import easyocr
```

--------------------------------

### Sample reader.readtext output

Source: https://www.jaided.ai/easyocr/tutorial

Shows the standard output format of reader.readtext: a list of tuples with text box coordinates, text, and confidence.

```python
[([[448, 111], [917, 111], [917, 243], [448, 243]], '高鐵左營站', 0.9247),
 ([[454, 214], [629, 214], [629, 290], [454, 290]], 'HSR', 0.9931),
 ([[664, 222], [925, 222], [925, 302], [664, 302]], 'Station', 0.3260),
 ...]
```

--------------------------------

### Use EasyOCR to read text from an image

Source: https://www.jaided.ai/easyocr

Import EasyOCR, create a Reader for Chinese simplified and English, and read text from an image file.

```python
import easyocr
reader = easyocr.Reader(['ch_sim','en'])
result = reader.readtext('image.jpg')
```

--------------------------------

### recognize method

Source: https://www.jaided.ai/easyocr/documentation

Recognize characters from text boxes. If no boxes are given, the whole image is treated as one text box.

```APIDOC
## recognize method

### Description
Recognize characters from text boxes. If no boxes are given, the whole image is treated as one text box.

### Parameters
- **image** (string, numpy array, or bytes) - Required - Input image.
- **horizontal_list** (list, default `None`) - Optional - See output of `detect`.
- **free_list** (list, default `None`) - Optional - See output of `detect`.
- **decoder** - Optional - Same as `readtext`.
- **beamWidth** - Optional - Same as `readtext`.
- **batch_size** - Optional - Same as `readtext`.
- **workers** - Optional - Same as `readtext`.
- **allowlist** - Optional - Same as `readtext`.
- **blocklist** - Optional - Same as `readtext`.
- **detail** - Optional - Same as `readtext`.
- **paragraph** - Optional - Same as `readtext`.
- **contrast_ths** - Optional - Same as `readtext`.
- **adjust_contrast** - Optional - Same as `readtext`.

### Returns
List of results.
```

--------------------------------

### Create EasyOCR Reader for Chinese Traditional and English

Source: https://www.jaided.ai/easyocr/tutorial

Initializes the EasyOCR Reader to recognize Traditional Chinese and English by passing a list of language codes.

```python
reader = easyocr.Reader(['ch_tra', 'en'])
```

--------------------------------

### Read from numpy array, bytes, and URL

Source: https://www.jaided.ai/easyocr/tutorial

Demonstrates how to pass a numpy array (from OpenCV), a bytes object, or a URL to readtext.

```python
img = cv2.imread('chinese_tra.jpg')
result = reader.readtext(img)

with open("chinese_tra.jpg", "rb") as f:
    img = f.read()
result = reader.readtext(img)

result = reader.readtext('https://example.com/chinese_tra.jpg')
```

--------------------------------

### detect method

Source: https://www.jaided.ai/easyocr/documentation

Detect text boxes only. Returns horizontal and free-form bounding boxes.

```APIDOC
## detect method

### Description
Detect text boxes only. Returns horizontal and free-form bounding boxes.

### Parameters
- **image** (string, numpy array, or bytes) - Required - Input image.
- **min_size** - Optional - Same as `readtext`.
- **text_threshold** - Optional - Same as `readtext`.
- **low_text** - Optional - Same as `readtext`.
- **link_threshold** - Optional - Same as `readtext`.
- **canvas_size** - Optional - Same as `readtext`.
- **mag_ratio** - Optional - Same as `readtext`.
- **slope_ths** - Optional - Same as `readtext`.
- **ycenter_ths** - Optional - Same as `readtext`.
- **height_ths** - Optional - Same as `readtext`.
- **width_ths** - Optional - Same as `readtext`.
- **add_margin** - Optional - Same as `readtext`.
- **optimal_num_chars** (int, default `None`) - Optional - If specified, boxes with estimated number of characters near this value are returned first.

### Returns
`horizontal_list, free_list`. `horizontal_list` is a list of rectangular boxes `[x_min, x_max, y_min, y_max]`. `free_list` is a list of free-form boxes `[[x1,y1],[x2,y2],[x3,y3],[x4,y4]]`.
```

--------------------------------

### Reader class

Source: https://www.jaided.ai/easyocr/documentation

Base class for EasyOCR. Initializes the OCR engine with language support, GPU usage, model storage, and network configuration.

```APIDOC
## Reader class

### Description
Base class for EasyOCR. Initializes the OCR engine with language support, GPU usage, model storage, and network configuration.

### Parameters
- **lang_list** (list) - Required - List of language codes to recognize, e.g., `['ch_sim','en']`.
- **gpu** (bool or string, default `True`) - Optional - Enable GPU.
- **model_storage_directory** (string, default `None`) - Optional - Path to directory for model data. If not specified, models are read from `EASYOCR_MODULE_PATH` (preferred), `MODULE_PATH`, or `~/.EasyOCR/`.
- **download_enabled** (bool, default `True`) - Optional - Enable automatic model download.
- **user_network_directory** (string, default `None`) - Optional - Path to user-defined recognition network. Default: `MODULE_PATH + '/user_network'`.
- **recog_network** (string, default `'standard'`) - Optional - Name of recognition network.
- **detector** (bool, default `True`) - Optional - Load detection model into memory.
- **recognizer** (bool, default `True`) - Optional - Load recognition model into memory.

### Attributes
- **lang_char** - All available characters in the current model.
```

--------------------------------

### Paragraph mode with paragraph=True

Source: https://www.jaided.ai/easyocr/tutorial

The paragraph=True option combines recognized text into paragraphs, grouping them into lines.

```python
reader.readtext('chinese_tra.jpg', detail=0, paragraph=True)
# ['高鐵左營站 HSR Station 汽車臨停接送區 Car Kiss and Ride']
```

--------------------------------

### readtext method

Source: https://www.jaided.ai/easyocr/documentation

Main method for a Reader object. Performs text detection and recognition on an image, returning a list of results.

```APIDOC
## readtext method

### Description
Main method for a Reader object. Performs text detection and recognition on an image, returning a list of results.

### Parameters
#### General
- **image** (string, numpy array, or bytes) - Required - Input image.
- **decoder** (string, default `'greedy'`) - Optional - Options: `'greedy'`, `'beamsearch'`, `'wordbeamsearch'`.
- **beamWidth** (int, default `5`) - Optional - How many beams to keep when using beam search.
- **batch_size** (int, default `1`) - Optional - Larger batch is faster but uses more memory.
- **workers** (int, default `0`) - Optional - Number of dataloader threads.
- **allowlist** (string) - Optional - Force EasyOCR to recognize only a subset of characters.
- **blocklist** (string) - Optional - Block a subset of characters. Ignored if `allowlist` is given.
- **detail** (int, default `1`) - Optional - Set to `0` for simple output.
- **paragraph** (bool, default `False`) - Optional - Combine result into paragraphs.
- **min_size** (int, default `10`) - Optional - Filter text boxes smaller than this (pixels).
- **rotation_info** (list, default `None`) - Optional - Rotate each text box and keep the best result. E.g., `[90, 180, 270]`.

#### Contrast
- **contrast_ths** (float, default `0.1`) - Optional - Low-contrast boxes are processed twice.
- **adjust_contrast** (float, default `0.5`) - Optional - Target contrast for low-contrast boxes.

#### Text Detection (from CRAFT)
- **text_threshold** (float, default `0.7`) - Optional - Text confidence threshold.
- **low_text** (float, default `0.4`) - Optional - Text low-bound score.
- **link_threshold** (float, default `0.4`) - Optional - Link confidence threshold.
- **canvas_size** (int, default `2560`) - Optional - Maximum image size.
- **mag_ratio** (float, default `1`) - Optional - Image magnification ratio.

#### Bounding Box Merging
- **slope_ths** (float, default `0.1`) - Optional - Max slope for merging.
- **ycenter_ths** (float, default `0.5`) - Optional - Max shift in y direction.
- **height_ths** (float, default `0.5`) - Optional - Max difference in box height.
- **width_ths** (float, default `0.5`) - Optional - Max horizontal distance to merge boxes.
- **add_margin** (float, default `0.1`) - Optional - Extend bounding boxes in all directions. Important for complex scripts like Thai.
- **x_ths** (float, default `1.0`) - Optional - Max horizontal distance to merge text boxes when `paragraph=True`.
- **y_ths** (float, default `0.5`) - Optional - Max vertical distance to merge text boxes when `paragraph=True`.

### Returns
List of results.
```

--------------------------------

### Read text from image file using reader.readtext

Source: https://www.jaided.ai/easyocr/tutorial

Reads text from an image file using the reader's readtext method.

```python
result = reader.readtext('chinese_tra.jpg')
```
Tokens:
Raw