Context-Aware Image Annotation and Retrieval on Mobile Device

In this paper, we present a novel method for image annotation and retrieval on mobile device by using contextual information. Image annotation is an effective way for content based image retrieval. However, manual annotation is an expensive and time consuming work, especially for mobile device. Here, we propose a semi-automatic way to give image annotation on mobile device. There are rich contexts for a mobile device, such as photo captured context, personal context and social network context. We synthesize these contexts and get useful semantic content of the photo. The synthesized results are treated as annotation suggestions. Our annotations include time, location, event, persons etc. We have implemented the annotation method in a prototype. The results show the method is simple and efficient.