Since technology is no longer restricted to the home or office, users are now capable of accessing information in virtually any sort of public setting. This creates situations where sensitive information, both organizational and personal in nature, can be seen and captured by nearby people and technology. Therefore, methods are necessary to ensure the privacy and security of information displayed on mobile devices. The authors have developed privacy blinders for mobile device displays that hide sensitive information from view while leaving other displayed information unobscured. Two experiments were conducted to test the usability of privacy blinders and their effect on information retrieval tasks. Results support the overall usefulness of the privacy blinder concept, and form a foundation for continued development of the technique.

    This project is just starting up and looking for one or more student(s) who would like to work on design, implementation and testing. The title indicates two different projects but the underlying techniques are the same.

    "Prosody" refers to the stress, rhythm, intonation, and pause structure in the speech signal. We want to generate text where certain aspects of the prosody are reflected in the shape of the text, e.g. volume might be indicated by the size of the letters. Potential applications of the work include enhanced text for young readers or foreign speakers, and automatic, interesting karaoke displays.

    For lip-synching, we wish to move animated lips in synch with digitized speech. Microsoft agents, for example, can do limited lip-syncing with speech generated from text. The goal here is to have these or other agents produce pre-recorded speech with proper lip movements and possibly other appropriate gestures.

    To automate both these tasks, we first extract word and phoneme boundaries from the recorded speech by matching it with the text in an ordinary transcription of the recording. I will demonstrate LISET, Microsoft's Linguistic Information Sound Editing Tool and then show a tool I built that dissects LISET's output (.lwv) files to collect the word and phoneme boundaries.