Optical Character Recognition

Zia Optical Character Recognition electronically detects textual characters in images or digital documents, and converts them into machine-encoded text. Zia OCR can recognize text in nine international languages and 10 Indian languages. You can check the list of languages and language codes from the API documentation.

Note:Catalyst does not store any of the files you upload in its systems. The files you upload are used for one-time processing only. They are not used for ML model training purposes either. Catalyst components are fully compliant with all applicable data protection and privacy laws.

You must specify the path to the image or document file that needs to be processed for OCR as a parameter to the open() method. This opens the file and returns a file object as a response.

Allowed file formats: .jpg, .jpeg, .png, .tiff, .bmp, .pdf

File size limit: 20 MB

You must pass the file path, model type, and languages as arguments to the extract_optical_characters() method. However, the model type and language values are optional. By default, it is passed as the OCR model type, and the languages are automatically detected if they are not specified.

The zia reference used below is already defined in the component instance page.

#OCR Implementation img = open('sample.jpg', 'rb') result = zia.extract_optical_characters(img, {'language': 'eng', 'modelType': 'OCR'})

A sample response is shown below :

{ "confidence":95, "text":"This is a lot of 12 point text to test the\nocr code and see if it works on all types\nof file format\n\nThe quick brown dog jumped over the\nlazy fox. The quick brown dog jumped\nover the lazy fox. The quick brown dog\njumped over the lazy fox. The quick\nbrown dog jumped over the lazy fox" }

Last Updated 2023-12-18 16:20:08 +0530 +0530