Memex Poland conference on the 6th of October 2022 brought together technology experts and enthusiasts. They could share their knowledge and predictions on future technological innovations and development directions. In his presentation, Karol Duzinkiewicz, the technical lead of the DAC.digital‘s project for the eye-tracking gaze estimation tool, gave insights into the state of eye-tracking, a technology for understanding users’ digital experience.
At the end of this summary, you will find a video recording of the presentation.
What are gaze estimation and eye tracking?
In the introduction of his presentation, Karol explains the basics of eye tracking and computer vision. The basic principle of this technology is to track where at which point the screen the user is looking at a given time. It allows for measuring eye positions and analysing where the attention goes and how long the person is looking at a given point on a surface (also referred to as a gaze point).
The idea has already been present for some time, and, in most cases, collecting the data and achieving the results required dedicated hardware or eyewear. As these devices are accurate, there are some disadvantages to using these computer vision systems. First, they were built in a way that isn’t comfortable, as they were chunky pieces of technology. These eye gaze estimation devices are expensive and, therefore, not easy to obtain. But most importantly, they are based on infrared technology, so they can’t be used outside.
Eye-tracking system for gaze tracking “in the wild”
In his presentation, Karol highlights the importance of comfortably using eye tracking and computer vision in the user’s natural environment for more practical eye-tracing applications. The ideal device would have a built-in camera and be easily accessible. Nowadays, most people own a smartphone, laptop or tablet equipped with a camera. Thus, using a personal device for eye-tracking research would be affordable and relatively easy to execute.
A significant milestone on the road to tracking eye movements and collecting eye-tracking data efficiently is ensuring that it can be done in a natural environment, whether enclosed or open, regardless of the lighting conditions.
Computer vision technology and practical eye-tacking use
There are many ways to practically apply eye tracking and computer vision in its broader spectrum. In his presentation, Karol mentions controlling devices without using hands or hand gestures instead of controllers like a mouse in a more comprehensive computer vision approach.
Eye tracking can contribute to automotive safety features to, for example, enable the car to detect if the driver focuses on the road or their phone. Another example is a more entertaining aspect of eye tracking and computer vision, virtual reality and other elements in the gaming industry, where the computer vision field has wide application prospects.
From a business point of view, eye tracking can be a valuable tool for conducting market research, which you could hear about in another Memex presentation by Florian Passlick. Karol’s presentation also mentions the project DAC.digital is working on together with eye square. This technology uses user image data scrolling commercial content on their phones to determine what interests them.
These are a few examples of adapting computer vision and eye tracking for different purposes and business applications. The possibilities are immense, and there are many still to discover.
Watch the video recording
Check out the video below to listen to the entire presentation.