Go Back to Insights Page

articles

AI Object Recognition and Metadata Enhancement

The integration of Object Recognition in machine learning helps AI’s capabilities to grow, as it’s used to understand and recognize visual content.

mog-insights-articles-AI-Object-Recognition-and-Metadata-Enhancement.png

The integration of Object Recognition in machine learning helps AI’s capabilities to grow, as it’s used to understand and recognize visual content.

Computer Vision aims to teach a computer to understand the content of a scene from an image or video.
Despite being a simple task for a human, understanding what is in an image is complex for computer software.
The growth of AI brought us to think about the possibilities of applying these techniques to professional video workflows.
Using computer vision to convert the video “itself” into metadata that can be easily searched and cataloged.

So, what’s the scenario for Object Recognition right now?

Imagine you have a library of hundreds of hours of video and you want to know what appears in each file and when.
If you did not catalog the file with that metadata and locators this task will require someone to watch the entire collection…
MAM4PRO can do this for you with AI.

ai-object-recorgnition-and-metadata-enhancment-initial-setup

ai-object-recorgnition-and-metadata-enhancment-initial-player


How can MAM4PRO do that for you?

You will only need to enable it in the MAM4PRO general settings.
In addition to the activation of Metadata extraction, Thumbnails, and Video Preview, there also be the possibility of enabling AI.
Once this feature is enabled, when an asset is detected, if there is a metadata sidecar with the detected object’s description, it will be loaded and previewed each time the asset is played.
The detected objects will be outlined with boxes that will contain the name of the object to which they refer.
In the player’s timeline, it will also be marked the position where there are detected objects.


When ingesting new content, this feature may be enabled for each workflow profile in its advanced options. By enabling the Artificial Intelligence Metadata Enhancement option, during the processing chain, the media will be analyzed and a sidecar file will be created with the detected object’s description.

As you can see, this is the current state-of-the-art of AI object recognition implementations on MOG solution MAM4PRO.
Developers will keep working on improving machine learning, and Artificial Intelligence will get smarter, and advanced as time goes by.
With MOG’s AI Object Recognition, you can simplify and automate metadata logging, making use of locators for better usability.
Still not convinced? Ask us for a demo!

AUTHOR

MOG

Posted on 2019-10-18

SHARE THIS ARTICLE

Let’s build something amazing, together.