Integrating computer vision techniques into a touch pad system

A key strength of touchpads, such as iPads or Galaxy Tabs, is that they provide portable access to the Internet and many applications that entertain and help managing the lives of users. The integration of computer vision methods into touchpads results in even more powerful devices that enable natural human-computer interaction. This thesis proposes two techniques of incorporating computer vision methods -- one technique supports touch-based interaction for biomedical image analysis, the other camera-based interaction for music therapy and entertainment: I'mCell is an application for annotating objects in images, for example, cells in phase-contrast microscopy images. MusicTracks recognizes a user's facial expression, captured by the camera of the touchpad, and plays music according to the user's mood. The I'mCell and MusicTracks applications have been implemented for the iPad. Users who experimented with the applications report them to be convenient because they enable efficient (I'mCell) and enjoyable (MusicTracks) interactions and are easy-to-use and portable