Image based Touristic Monument Classification using Graph based Visual Saliency and Scale-Invariant Feature Transform

Grigorios E. Kalliatakis, Tsampikos Kounalakis, Georgios Papadourakis, and Georgios A. Triantafyllidis


SIFT, Graph Based Visual Saliency, Image Classification


This paper presents an image-based application using Graph Based Visual Saliency (GBVS) and Scale-Invariant Feature Transform (SIFT), aiming at simple image classification of well-known touristic monuments in the geographic area of Heraklion, Crete, Greece. For this purpose, photographs taken at various sites of interest are being compared to an existing database containing photos of these sites at different angles and zoom. The time required in such application is an important element. To this goal, the proposed application employs SIFT algorithm to compare the user-taken photographs with the database photographs, that have been previously processed according to the Graph Based Visual Saliency technique, in order to minimize the “noise” of the monument’s background and keep only the SIFT features that will help faster and more accurate classification. The application is then able to classify these photographs fast, helping the user to better understand what he sees and in which area he had this photograph.

Important Links:

Go Back