GazeHits is a design assistant tool to know how your design is perceived in terms of visual attention. It makes easy to analyze alternative designs and decide when a work is ready to present it to your customer.
It bases itself on state-of-the-art visual attention computer models to identify the elements and features that captures attention. It decomposes images in a similar manner to neural responses observed in the visual cortex, applying operations that take place all along the visual pathway (e.g. decorrelation, variance normalization).
Visual attention is the result of combining three main factors: saliency due to the physics of the scene, relevance related to the intentions and knowledge of the observer, and biases of the human visual system (e.g. the tendency to gaze the center).
Saliency is clearly the most interesting attention driver for design purposes. You cannot modify biases of the human visual system and you usually cannot be sure of the intentions and the experience of your observer, but you may modify a color or a texture to catch attention by boosting saliency.
Visual attention analysis on a street context.
GazeHits helps you in a variety of situations in which you need to guess how attention will be driven under free viewing assumptions. For example, it can be used to assess performance of visual designs like ads on their context, because it helps to understand how the context will impact on the attention of its recipients.
GazeHits is a project by:
We aim to deliver effective tools for visual perception analytics relying on ultimate computational models of visual attention and perception as core technology.
GazeHits is still in beta testing but it is continuously evolving to provide the best tools and features for attention analisis. We are glad to offer experimental features to collaborators. Contact us if you are keen to collaborate.
Do you have an application scenario that require some specific features? Just drop an email to email@example.com. We are open to include functions that match our roadmap.