CoolFace
Apppublic

aravindh-s/multiocr

sourceHugging Facemitupdated 3y agoView on Hugging Face
1likes
App README

Multiocr

This package intends to give a common interface for multiple ocr backends

Installation

pip install multiocr

Supported OCR Backends

  • —Tesseract
  • —PaddleOCR
  • —Aws Textract
  • —EasyOCR
  • —DocTR OCR

the output for all ocr backend will be simillar

Code Example

Tesseract

python
from multiocr import OcrEngine

config = {
    "lang": "eng",
    "config" : "--psm 6"   
}
image_file = "path/to/image.jpg"
engine = OcrEngine("tesseract", config)
text_dict = engine.text_extraction(image_file)
json = engine.text_extraction_to_json(text_dict)
df = engine.text_extraction_to_df(text_dict)
plain_text = engine.extract_plain_text(text_dict)

PaddleOCR

python
from multiocr import OcrEngine

config = {
        "lang":"en"
    }
image_file = "path/to/image.jpg"
engine = OcrEngine("paddle_ocr", config)
text_dict = engine.text_extraction(image_file)
json = engine.text_extraction_to_json(text_dict)
df = engine.text_extraction_to_df(text_dict)
plain_text = engine.extract_plain_text(text_dict)

Aws Textract

python
from multiocr import OcrEngine

config = {
    "region_name":os.getenv("region_name"),
    "aws_access_key_id":os.getenv("aws_access_key_id"),
    "aws_secret_access_key":os.getenv("aws_secret_access_key")
}
image_file = "path/to/image.jpg"

engine = OcrEngine("aws_textract", config)
text_dict = engine.text_extraction(image_file)
json = engine.text_extraction_to_json(text_dict)
df = engine.text_extraction_to_df(text_dict)
plain_text = engine.extract_plain_text(text_dict)

EasyOCR

python
from multiocr import OcrEngine

config = {
    "lang_list": ["en"]
}
image_file = "path/to/image.jpg"
engine = OcrEngine("easy_ocr", config)
text_dict = engine.text_extraction(image_file)
json = engine.text_extraction_to_json(text_dict)
df = engine.text_extraction_to_df(text_dict)
plain_text = engine.extract_plain_text(text_dict)
print()

if you want to access the output of each individual ocr engine in their own raw format, we can fetch it this way

raw_ocr_output = engine.engine.raw_ocr

config is the each ocr's input parameters and it should be python dictionary. if not given, it'll default to each respective libraries default parameters

the input parameters for each ocr differs, and you can look at its respective repo for all allowable parameters

Reference & Acknowlegements

AWS Textract will not work because it requires access_key and secret