Scene classification is a fundamental problem in image understanding. Scene classification has a high potential for improving the performance of other computer vision applications such as browsing, retrieval and object recognition. However, scene classification is not an easy task owing to their variability, ambiguity, and the wide range of illumination and scale conditions that may apply. A number of works have been proposed for scene classification. However, it lacks a quantitative comparison. This paper has two mains contributions. Firstly, the authors provide an analysis of different types of feature for scene representation as well as a quantitative comparison of these features for scene classification. Secondly, the authors present a new application of scene classification: advertising service based on image content.