Thoughtsort is part of a prototype framework which aims to bring eye tracking to the web and enable developers to create dynamic experiences with a wider awareness of the viewing user.
An example application using this frameworks is thoughtsort which provides adaptive image searching with the identification of images being based on the users gaze path.
In a traditional image search the content remains static with only the initial keyword term being used to gather a selection of images. Thoughtsort overcomes this by dynamically gathering new content and retuning it back into your results based on your interests as detected by your gaze path. This allows you to scroll through the images much like a time line with thoughtsort automatically adjusting the results to find similar content to that which you have expressed an interest in.
The framework was originally developed around the opengazer eye tracking project and so currently only operates on linux based systems. Due to the libraries used it is unsuitable for unassisted installation and so this version will not be released until a later date.
A new Adobe Flash based version is currently being developed which will improve upon the initial prototype and allow cross platform support. This version is currently unable to process your gaze path but it's progress can be checked at: