Article Preview
TopIntroduction
The rapid development in digital cameras, GPS-equipped cameras, smart phones and other advanced devices, created an interest among people to capture and collect more photographs. This lead to the availability or increase in multimedia collections. To share this information, social media sites such as Facebook, YouTube and Flickr created a fine opening. As an outcome Becker et al. (2009), details that these social media sites hold more user-thrown materials of photos, videos, textual content, etc., and also lead to a mounting record of the culture and shared experience. No one is ready to spend time to organize photographs which leads to many being untouched. This created the need for image indexing and retrieval. Sandhaus et al. (2011) styles the lack in the social aspect of image retrieval and the need for semantic photo understanding. Contextual information, given by Tankoyeu et al. (2011), reports the environmental condition of an image and the basic property of an event it belongs to. This information adds semantics to the photographs. The three image context explained by Tankoyeu et al. (2011) includes, first the photo parametrical context which accompanies the moment of image capturing, the timestamp, GPS coordinates, device information, lens information and flash information. The second is the image environmental context involving the season of the year, weather information and place information. The third is the user-generated context, containing textual description, tags and comments in the social networks.
Tags are described by Rattenbury et al. (2007) as freely-wished and positioned, short list of words or textual labels given by the user for these resources. These tags are very powerful as they hold semantics, state the resource, useful feature in social media and help extracts knowledge out of them. Van Laere et al. (2013) used tags from Flickr photos, to characterize places of interest. With the availability of GPS-enabled capture devices, geographic information is automatically embedded especially in photos. As a result, there is an increase in the number of geo-referenced photos available online. The geo-referenced multimedia is community contributed and describe various aspects of people’s life such as events, places, activities, etc. Zheng et al. (2011) highlights the numerous location-based services being introduced in the market. Among them foursquare is popular. A thorough location and event information have spawned an enormous amount in such social services. Most popular locations such as tourist places and any location of interest, shops, restaurants, etc., can be identified using location-based services.
A set of visible activities can be naturally expressed as events. According to Firan et al. (2010), “event is a specific thing happening at a specific time and place”. Everyday doings, meeting, etc., of an individual represent events. Events vary widely, ranging from planned, known occurrences such as a concert or a parade, to spontaneous, unplanned incidents such as an earthquake or death of a celebrity. An event can occur for some two to three hours, a day, several days, a week, several weeks and months, so Rabbath et al. (2011) explains that events vary temporally. Every human, comes across several significant activities like devotional meetings, birthdays, climate change, technical meetings, friends get together, earthquakes, public protests, tsunami, sports matches, etc. There are two different aspects to events according to Firan et al. (2010), one representing local events and the other global events. Small, personal experiences such as vacations, celebrations, birthdays, meetings in offices, get-togethers in work environment, etc., represent local events and the world known activities such as sports matches, disasters, public protests, medical achievements, etc., represent global events.