Keep an eye on your photos: automatic image tagging on mobile devices

Abstract

In this paper we present how to tag images automatically based on the image and sensor data from a mobile device. We developed a system that computes low-level tags using the image itself and meta data. Based on these tags and previous user tags we learn high-level tags. With a client-server-implementation we source out computational expensive algorithms to recommend the tags as fast as possible. We show what are the best feature extraction methods in combination with a machine learning technique to recommend good tags.