Which service is used to extract text from scanned documents, such as old newspapers?

Amazon Textract is designed specifically for extracting text and data from scanned documents, such as digitized versions of old newspapers, forms, and tables. It employs advanced machine learning techniques to analyze the layout, structure, and content of documents, enabling it to accurately recognize printed text and even handwriting. This service goes beyond simple optical character recognition (OCR) by not only detecting text but also understanding the relationships between different elements within a document, which is crucial for comprehensively extracting information.

In contrast, Amazon Rekognition focuses on image and video analysis, including object and scene detection, facial recognition, and activity recognition, making it unsuitable for text extraction from documents. Amazon Transcribe is oriented towards converting spoken language into text, ideal for transcription of audio files rather than for written documents. Lastly, Amazon Translate is a machine translation service used for converting text from one language to another, rather than for extracting text from images or scanned documents.