Leveraging machine learning to improve privacy control
Technology
February 4th 2021

Leveraging machine learning to improve privacy control

Amy Griffith
Amy Griffith

The team at MODRON have been working on solving an important problem with a very interesting approach.

We have come across a number of scenarios where certain participants in a dispute resolution process would have a need to hide the detail of their room or space while on a video call. In response to this we have just released a new feature we are calling Privacy Blur.

This new feature enables participants on a video call to instantly blur their background in real time. Importantly this is achieved through your Spaces account without having to install anything on your device.

Robust, realtime face and head-tracking within MODRON Spaces
Robust, realtime face and head-tracking. Our Privacy Blur feature detects more than 68 facial feature points at up to 60 frames per second.

To achieve Privacy Blur we needed to be able to reliably detect faces and facial features in real time. Using a variety of proprietary data models, and also advanced machine learning in 3D we are able to provide extremely fast and precise face detection, fast and robust eye, nose and chin detection in real time during Space Calls.

Our Privacy Blur feature is able to anonymously track any face under almost light conditions from any angle, even when the head is turned 180 degrees either side. This allows us to then recognise which part of the video image is the background that should be blurred leaving the persons head unchanged and visible. This process occurs continuously for every frame of the video call.

Robust, realtime face and head-tracking within MODRON Spaces
Our Privacy Blur feature can track any face under almost light conditions from any angle, even when the head is turned 180 degrees either side.

We are only able to achieve these results by leveraging the power of machine learning. After analysing countless images of people in different rooms the system can be guided to use this data to become increasingly effective and accurate at recognising the faces of people from the rest of a video image in almost any scenario. Pretty cool and exciting stuff I think!

This is just one of the projects we are working on that leverages the power of machine learning and artifical intelligence. Stay tuned for more exciting releases and updates just around the corner from MODRON.

Using Privacy Blur guide

Partner with MODRON to power your dispute resolution processes

MODRON wants to support people and organisations in the world who are working to resolve disputes of any and every type. Please contact us to discuss how we can support your efforts. hello@modron.com

Share:

Check out some other articles

Using recorded messages to improve access to justice

Using recorded messages to improve access to justice

We have seen simple but extremely effective uses of our recorded message tools in addressing some of the Access to Justice principles, specifically around accessibility. If for example, one or more parties have literacy or language difficulties...

Read Story
It's not just about video calls in dispute resolution. Here's why

It's not just about video calls in dispute resolution. Here's why

Video calls are a core component of effective online dispute resolution as they allow free flowing conversations to take place from anywhere. But there are so many other tools that can and should be leveraged in employing best practice online dispute resolution.

Read Story
Space Calls vs breakout rooms. What's better suited to disputes?

Space Calls vs breakout rooms. What's better suited to disputes?

We designed Space Calls to be the perfect video call solution for dispute resolution. Are Space Calls better suited to dispute resolution then other options such as Zoom's breakout room functionality?

Read Story
Icon