Almost all tourism search providers are restricted to a text-based or purely attribute-based search for points of interest (POI). In such cases users must know their search criteria and needs before seeking any kind of touristic object (e.g. a destination of the upcoming trip). However, this does not correlate with the user’s real search behavior – especially in early stages of the travel decision.
To solve this problem and to support users in their early trip-planning phases, we developed an image-based search engine called PixMeAway. By the help of a Web-based portal and a pre-defined set of pictures, we capture the (implicitly given) emotions and desires of the user. The users can select those pictures that reflect their current emotional state and as a result, the platform recommends appropriate tourist objects (e.g. a city, a site or even an activity, etc.). Our approach combines the picture search with a recommender system and community aspects, which makes it appropriate for other sectors and not only for tourism.
PixMeAway was developed in Java, Apache Wicket, Solr, MySQL