You can design by only watching!

So far, in an interactive evolutionary computation system, it was necessary to input the user's evaluation into the system. However, we think that if the system can understand the user's preferences and generate the user’s favorite design without inputting evaluation into the system by the user, it is a breakthrough.

Gaze information is human biological information. When a human being receives stimulation of something, its reaction appears in every part of the body without being conscious. For example, when you meet a celebrity, you are excited. Therefore, your pulse rise. From this point of view, there is the word “Sometimes the eyes say more than the mouth.” Therefore, it can hypothesize that the potential preference of the human can be known from the human gaze. In our research, based on this hypothesis, we are developing an interactive evolutionary computation system with gaze information.

The figure below is a summary of the proposed system. User's evaluation behavior is replaced with gaze information. In short, the user only looks at the design.


Overview of the proposed system

To measure the line of sight, we need a gaze measuring device. In our laboratory, using Omron's face vision and Tobii's gaze measuring device, we build various systems targeting a single user or multiple users. The figure below shows the scenery of experiments for single user and multiple users. Depending on a use, measurement device and experiment environment are changed. In the case of a single user, we measure the line of sight with an infrared device. In the case of multiple users, we measure the line of sight with image analysis.


Experiment scenery (for single user)

Experimental scenery (for multiple users)

As prototypes of these systems, we have proposed electronic catalog systems that reflect individual sensibilities and character coordination based on gaze information from multiple users. For more details, click here → Smart e-CatalogCoordinates with everyone!

Related our works
【Academic Papers】Hiroshi Takenouchi, Masataka Tokumaru, "Interactive Evolutionary Computation System with User Gaze Information", International Journal of Affective Engineering, Vol.18, No.3, pp.109-116, 2019-07.

▲pagetop