The next generation of NeuroVision has just been released with its recent engine update. This update includes a critical addition: the ability to detect faces in both images and videos.
What does it mean?
Humans are especially prone to see faces (even in things that are not faces). We tune in to faces almost instantly, and we can easily see if there is a face in an image or not. Now, the NeuroVision output mirrors this, in addition to the other cues for automatic attention.
How do we do it?
A completely new face detection engine has been adopted from state-of-the-art algorithms. The algorithm calculates the probability that a part of an image has one or more faces. It then marks where this face is. The more probable that it is a face, the warmer and broader the heat map from the center of the face.
What is the science behind it?
We have long known that faces automatically grab attention. Being social creatures as we are, we are automatically tuned to faces. But in all respects, face detection is occurring later in the visual process. This is one of the reasons that we have not included face detection until now. A second reason is that computational neuroscience has not produced a sufficiently reliable approach to detect faces. However, today, we know that faces are powerful eye-grabbers and that we can reliably detect faces with our algorithm. The gain to this even better prediction of hardware-enabled eye-tracking.
Do you need to do anything different when running your analyses?
Nope, it’s all automatically included in your heat map and fog map output
To try it out, go to your NeuroVision account (or register now).
All the best from the NeuroVision team!