This paper discusses and compares various approach to automatic landmark<br />recognition in pictures, based upon image content analysis and<br />classification. The paper first compares various visual features and<br />image similarity functions based on local features. Finally it discusses<br />and compares a new classification technique to decide the landmark<br />contained in an image that first classifies the local features of the<br />image and then uses this result in order to take a final decision on the<br />entire image. Experiments demonstrate this last approach is the most<br />effective one. The discussed techniques were used and tested in the<br />project VISITO Tuscany.
Amato, G., Falchi, F., & Rabitti, F. (2012). Landmark recognition in VISITO Tuscany. In Communications in Computer and Information Science (Vol. 247 CCIS, pp. 1–13). https://doi.org/10.1007/978-3-642-27978-2_1