Rewind: Leveraging Everyday Mobile Phones for Targeted Assisted Recall
This paper presents an assisted recall system, Rewind, that employs automatic image capture on mobile phones and filtering of images for end-user viewing. The usability of image-based assisted recall systems is limited by the large number of images through which the user must navigate. Rewind is a scalable system of everyday mobile phones and supporting web services that we developed to explore how client-and server-side image processing can be used to both lower bandwidth needs and streamline user navigation. It has been in use since August 2007 as part of a pilot study supervised by an epidemiologist to evaluate its utility for improving recall of dietary intake, as well as other shorter and more exploratory trials. While the system is designed to accommodate a range of image processing algorithms, in this first prototype we rely on simple filtering techniques to evaluate our system concept. We present performance results for a configuration in which the processing occurs on the server, and then compare this to processing on the mobile phones. Simple image processing on the phone can address the narrow-band and intermittent upload channels that characterize cellular infrastructure, while more sophisticated processing techniques can be implemented on the server to further reduce the number of images displayed to users.