Crowd-sourcing using images and eye-tracking

graphical summarisation of mass image feedback from crowds Description: A research project in the texture lab is investigating the use of images to feedback information to designers on the feelings from crowds. The crowd's task is simply to select images in response to questions such as "pick images which you associate with your 'loves' for this design", "pick images which you associate with your 'hates' for this design". The task is then to automatically form a visual precis (collage) of masses of images fed back from lots of individuals in such a way that the designer can understand at-a-glance the feelings of the crowd. The project would therefore concern the gathering of visual similarity information about images and using this to perform the clustering of images, that is used to produce the collage. A very interesting extension would be to use our eye-tracker so that users only to glance at images to provide feedback (instead of explicitly clicking on images in order to "choose" them).