AWS Document Understanding Reference App

Overview

This reference application helps understand the contents of general documents (PDFs, images, audio, video). These integrations can be easily embedded into any Appian application, not only to perform the task of understanding the context of the uploaded document but also to provide a seamless user experience.

Key Features & Functionality

• It uses AWS Textract to extract the text and line blocks from PDF or image files and utilizes AWS Comprehend to analyze the text for detecting key phrases.
• Additional Textract APIs for Identification Document Extraction and Querying PDF Documents.
• For audio (mp3) and video (mp4) media files, the app will employ AWS Transcribe to generate a text transcription from the media files.
• For image files, the app will use AWS Rekognition to detect key objects from the image. A sample use case is if a gun is detected in the image, actions will be taken accordingly.
• Comprehend NLP service to uncover insights and connections in text files.
• When the "End" button is clicked, the uploaded file will be deleted from both Appian and AWS S3.