Tools for image annotation Using context-awareness, NFC and image clustering

  • Randi Karlsen Department of Computer Science UiT The Arctic University of Norway, Tromsø, Norway
  • Anders Andersen Department of Computer Science UiT The Arctic University of Norway, Tromsø, Norway

Abstract

Annotation of images is crucial for enabling keyword based image
search. However, the enormous amount of available digital photos makes
manual annotation impractical, and requires methods for automatic
image annotation. This paper describes two complementary approaches
to automatic annotation of images depicting some public attraction. The
LoTagr system provides annotation information for already captured,
geo-positioned images, by selecting nearby, previously tagged images
from a source image collection, and subsequently collect the most
frequently used tags from these images. The NfcAnnotate system
enables annotation at image capture time, by using NFC (Near Field
Communication) and NFC information tags provided at the site of
the attraction. NfcAnnotate enables clustering of topically related
images, which makes it possible annotate a set of images in one
annotation operation. In cases when NFC information tags are not
available, NfcAnnotate image clustering can be combined with LoTagr
to conveniently annotate every image in the cluster in a single operation.

Section
Artikler