Automatic hotel photo quality assessment based on visual features

Photos have been an important means to transfer destination image to potential travelers, as they are widely available on travel related web sites. However, there has been very limited attempt to assess visual quality of online photos, which plays an important role in influencing traveler emotion and their travel intention. This is probably due to the limited background in photography among tourism researchers, and the inefficiency of manual assessment approach. Aiming to overcome these barriers, this paper presents a computational approach to visual features extraction for automatic photo quality assessment. We describe a number of visual features, which are helpful in reflecting the photo quality, and then validate their performance through a number of experiments using a large-scale data set of online travel photos. The introduced approach has the potential to facilitate the evaluation of visual quality of photos on travel web sites and online travel photos.