Build 2022: Azure Video Indexer Enhancements

by | May 27, 2022 | Technology | 0 comments

This article is contributed. See the original author and article here.

At MS Build 2022, Azure Video Indexer service is becoming the primary video analytics offering Applied AI services, a GA service for analyzing stored videos (e.g. media libraries), with a single endpoint,  more than 35 video and audio AIs – mainly based on Azure Cognitive Services. 


 


“Hi, My name is Azure Video Indexer”


 


Azure Video Indexer (VI), formerly known as Azure Video Analyzer for Media, was recently renamed back to VI, avoiding confusion for our media customers.  VI is officially the primary applied AI service for videos. See Applied AI Azure page (in the bottom)._ VI has a new product page in ACOM and a new pricing page in ACOM; You can explore the available regions in regions availability Azure page and to learn about VI APIs in Azure Video Indexer developer portal.  Our web experience has a fresh landing page, and the product documentation was also up to date with recent changes. In Azure portal, the resource provider name from now on is “Video Indexer.” As of May 1st, our new updated repository of Azure Video Indexer widget was renamed. Use https://www.npmjs.com/package/@azure/video-indexer-widgets instead.


Oh, and we are here to stay with this name. 


 


insagiv_0-1653582923082.png


Figure #1, Azure Video Indexer Web experience 


 


Our focus recently as a GA service, was supportability and quality of service. 


 


We are thrilled to announce the following enhancements: 


 


Better quality of AIs 


 


Line breaking in transcripts improvements 


Improved line break logic to better split transcript into sentences. New editing capabilities are now available through the Azure Video Indexer portal, such as adding a new line and editing the line’s timestamp. For more information, see Insert or remove transcript lines. We have invested in quality improvements of our audio pipeline so we expect much better results starting May 1st. 


 


Video Insights improvements


Object Character Reader (OCR) quality is improved by 60%. Face Detection quality is improved by 20%. Label accuracy is improved by 30% over a wide variety of videos. These improvements are available immediately in all regions and do not require any changes by the customer.


 


Closed Captioning files now support including speakers’ attributes


Azure Video Indexer enables you to include speakers’ characteristic based on a closed captioning file that you choose to download. To include the speakers’ attributes, select Downloads -> Closed Captions -> choose the closed captioning downloadable file format (SRT, VTT, TTML, TXT, or CSV) and check Include speakers checkbox.


 


Better Security and Supportability 


 


Service tag


Azure Video Indexer is now part of Network Service Tags. Video Indexer often needs to access other Azure resources (for example, Storage). If you secure your inbound traffic to your resources with a Network Security Group you can now select Video Indexer as part of the built-in Service Tags. This will simplify security management as we populate the Service Tag with our public IPs.


 


Azure Monitor integration


Azure Video Indexer now supports Diagnostics settings for Audit events. Logs of Audit events can now be exported through diagnostics settings to Azure Log Analytics, Storage, Event Hubs, or a third-party solution.


The additions enable easier access to analyze the data, monitor resource operation, and create automatically flows to act on an event. For more information, see Monitor Azure Video Indexer.


 


Celebrity recognition feature toggle


You can now enable or disable the celebrity recognition model on the account level (on classic account only). To turn on or off the model, go to the account settings > and toggle on/off the model. Once you disable the model, Video Indexer insights will not include the output of celebrity model and will not run the celebrity model pipeline in your account. 


 


You can explore the list of our recent features for NAB, 2022 in this blog post. 


 


New: Call to join our limited program: Run video and audio AIs in the browser 


 


At //Build 2022, in the 2nd day, Microsoft CTO Kevin Scott, shared how new advances in AI and other tech and software innovations are empowering developers and creators to do more, be more productive and creative, and enrich the lives of everyone on the planet in the The Future of AI Development Tools. 


 


Imagine how the experience would look like if any social network company who gives a channel for content creators to upload a video or audio files or to allow live video or audio streaming, would also allow a secure way to enrich and improve the content in the browser, before everybody is able to see it. Social networks would easily be able to e.g. moderate the content, suggest hashtags for the video content and allow dubbing for the content which is created making it more trusted and accessible. 


 


So, how can we ease the pain for a web developers, without asking them to explore multiple open sources, in a secured manner powered by Azure? 


Today we are announcing an exciting new limited-access program (private preview) we have incubated, which will be released first for internal use extending to all Microsoft owned companies. It’s temporary name is “On device AI” (ODAI) – innovative technology in the domain for content creation and social networks, that runs artificial intelligence in the browser in a low-code-low-cost manner


 


With ODAI extension, web developers working for social networks such as LinkedIn, can use Azure packaging service to generate a customized-encrypted package, with vision and/or speech AI pipeline of choice, in a secured manner, optimized for the social network business needs. The deployable package content runs seamlessly in the browser using generic existing web standards spanning from laptops to smartphones and tables, and allows to benefit from the client hardware to significantly reduce the provider costs, in this case the social network.


 


As mentioned, it’s a private preview program. If you wanna learn more, contact visupport@microsoft.com


insagiv_0-1653668674358.png


Figure #2, auto generate tags extracted from topics in the video, using in browser AI new technology for content creators and social networks 


 


Join Azure Video Indexer Today! Explore, Learn, Share


 


For those of you who are new to our technology, we encourage you to get started today with these helpful resources:



 

Brought to you by Dr. Ware, Microsoft Office 365 Silver Partner, Charleston SC.

Submit a Comment