AI Innovation Projects - Automation & Enrichment

Document created by harry.peek@alfresco.com Employee on Apr 9, 2018Last modified by harry.peek@alfresco.com Employee on Apr 12, 2018
Version 9Show Document
  • View in full screen mode

In this post we'll highlight projects using technologies to automate workflows and enrich content metadata, for example: Text analysis, entity extraction, sentiment analysis, image & video tagging. 

 

This will grow over time as more projects emerge from both Alfresco and the community. Got a project you'd like featured? Let me know!

 

 

Image & Text Analysis with AWS Rekognition & Comprehend

The Alfresco demo uses a fictional use case around an insurance company. As images or text files are uploaded into the system they are sent asynchronously to Rekognition or Comprehend for analysis, the content type of the node is changed to an appropriate custom type (acme:insuranceClaimImage or acme:insuranceClaimReport) and a unique ID generated.

If an image is detected Rekognition is used to look for a Cars, Motorcycles, Boats, Electronics, Jewelry, Wristwatches, Clocks, Bicycles, Sport Equipment and Furniture. If detected, the acme:claimType property is set appropriately otherwise the default value of "Unknown" is used.

If a text file is detected Comprehend is used to look for the claim adjusters name and the date and location of the visit within the text. The sentiment of the text is also analysed to determine whether a follow up visit is likely to be required. The extracted values are set on the acme:claimAdjuster, acme:visitDate, acme:visitLocation and acme:visitFollowUp properties, respectively.

Read more
Get the code

Process Services & the IoT

Nathan McMInn talks about using the AWS IoT platform with Alfresco Process Services and why it makes sense to do so.

 

 

Alfresco Photo Analyzer Module

David Antón's project from for the last Global Hack-a-thon which analyses photos of people and extracts information about their gender, age, face expressions, emotions and others. So you can make a bulk photo import to Alfresco, and then make searches (for example, to know how many 30 age people photos are in the repository).

Read more

Get the code

 

Integrating Alfresco with the new Google Vision API

Angel Borroy describes how they've integrated with the Google Vision & Translate APIs to add automatic image tagging to Alfresco.

Read more

Get the code

 

How to integrate cognitive services using ADF, CS and PS

Eugenio Romano shows how to build an ADF app with sentiment analysis using Microsoft Azure's services.

Read more

Get the code

 

Deep Learning Integration for Alfresco Media Content Analysis (Images & Videos)

Demo from Venzia integrating their deep learning server with Alfresco to analyse image and video content

Attachments

    Outcomes