CSE Developer Blog

Detecting “Action” and “Cut” in Archival Footage Using a Multi-model Computer Vision and Audio Approach with Azure Cognitive Services

Nile Wilson September 27, 2021 Sep 27, 2021 09/27/21

Movies and TV shows require multiple takes per scene and may have a substantial amount of archival footage as a result. Here, we use Azure Cognitive Services and custom code to develop a multi-model Machine Learning (ML) solution to automatically detect discardable footage to save media companies manual archiving hours and storage space.

Building an Action Detection Scoring Pipeline for Digital Dailies

Samuel Mendenhall September 20, 2021 Sep 20, 2021 09/20/21

Media companies capture footage filmed for the entire day in what's known as ‘digital dailies’. When talking about terabytes and petabytes of content, storage costs can be a factor. Lets explore Machine Learning approaches to identify which content can be archived or discarded which will save on those storage costs.

Enabling COVID-19 Exposure Notifications Across the United States on Azure

Andre Briggs September 7, 2021 Sep 7, 2021 09/7/21

During pandemics, the ability to let people know if they have been exposed to a contagious disease is essential for public health practitioners. This is known as contact tracing and the COVID-19 pandemic is no exception. In this post, I describe the journey of our Microsoft team to deploy an exposure notification server on Azure.

Entity Disambiguation Using Search Engine

Maysam Mokarian June 14, 2021 Jun 14, 2021 06/14/21

This blog post proposes a methodology to disambiguate misspelled entities by comparing the search retrieval performance with different custom search analyzers in a search engine.

Using Azure Cognitive Services to Analyse Evidence in Public Safety and Justice

Sonal Patil June 7, 2021 Jun 7, 2021 06/7/21

A scenario commonly encountered in public safety and justice is the need to collect and index digital data recovered from devices, so that investigating officers can perform evidence-based analysis. We recently built an advanced evidence analysis platform that uses Azure AI services for automated labelling of media and documents.

Observability for Event Stream Processing with Azure Functions, Event Hubs, and Application Insights

Shervyna Ruan May 13, 2021 May 13, 2021 05/13/21

With distributed tracing handled by Azure Functions, Azure Event Hubs, and Azure Application Insights behind the scenes, Azure Application Insights provides useful monitoring visualizations of the system that helped us easily understand system performances and troubleshoot failures.

Unit Testing Azure DevOps UI Extensions

Florian Wagner March 29, 2021 Mar 29, 2021 03/29/21

Mocking cloud service dependencies in modern applications is essential for independent unit testing. We show how to mock all relevant parts of the Azure DevOps Extension SDK, Azure DevOps Extension API as well as other external APIs to create meaningful unit tests for Azure DevOps UI Extensions and reach full code coverage.

Developing and Deploying a Recommender Model for Continuous Systematic Literature Reviews

Paolo Tenti February 10, 2021 Feb 10, 2021 02/10/21

This blog post is about using the Microsoft Academic Graph and NLP to build a personalized recommender system to suggest new scientific publications to researchers maintaining Systematic Literature Reviews.

Evaluation Framework for Information Extraction

Omri Mendels January 13, 2021 Jan 13, 2021 01/13/21

In this blog post we cover the process, requirements, and the design of an evaluation framework for NLP and Information Extraction. We cover the reasoning behind such a framework, and discuss its implementation with examples from a Named Entity Recognition evaluation point of view.

Building A Clinical Data Drift Monitoring System With Azure DevOps, Azure Databricks, And MLflow

Nile Wilson October 29, 2020 Oct 29, 2020 10/29/20

Hospitals around the world regularly work towards improving the health of their patients as well as ensuring there are enough resources available for patients awaiting care. During these unprecedented times with the COVID-19 pandemic, Intensive Care Units are having to make difficult decisions at a greater frequency to optimize patient health ...