Data entry is boring. We all hate it, but one team in Azure Cognitive Service decided we needed a better way that can take away the need to manually enter data. As part of a digital transformation, users can leverage an AI solution to automate and reduce the cost of converting documents (such as invoices and receipts) and forms into structured data for further processing. Azure Form Recognizer is an Azure Cognitive Service focused on using machine learning to identify and extract text, key-value pairs, and tables data from documents. Applications for Form Recognizer service can extend beyond just assisting with data entry. It could also be used in integrated solutions for optimizing the auditing needs of users, letting them make informed business decisions by learning from their expense trends or matching documents with digital records.

Let’s take an example to understand how a user might want to use Azure Form Recognizer to populate expense fields from receipt as part of an expense maintaining data entry app. This could be a web or mobile app using the Form Recognizer client library internally to interact with the service. The user of the app can provide a URL to a receipt or choose to upload file data for the document which they want to recognize expense field information. Let’s use a document URL for the examples below. The underlying client library then feeds this URL to the Form Recognizer service and outputs the relevant expense related information of the provided document.

Let’s see some code on how the app might want to use this library. Although, you can use any of our languages for this (.NETPythonJava, or JavaScript/TypeScript), we will be using Java for the examples today. These libraries are currently in preview and are supporting the v2.0-preview version of the Form Recognizer service.

#azure sdk #java #machine learning #azure

Introduce Form Recognizer Client Library: AI-powered Document Extraction
3.30 GEEK