The rapid development of the camera technology has affected the amount of photo captured. This development has contributed to camera sizes being integrated into many devices such as mobile phones, tablets, computers and more. Memory capacity for such devices has advanced as well allowing people to take lots of photos without limitations. However, as the size of the photo collections grows bigger, it becomes hard to retrieve a particular photo, unless they are organized, and are browsed, to allow users to identify them visually. Image retrieval has gain attention since 1970's. However, most of the available systems, to retrieve images, require users to manually annotate photos, which take a lot of time and effort. This paper presents a new method to annotate photos based on places automatically. Image matching technique is used to find similar photos related to the place from a geotagged database. Once a match is found from the geotagged database, a reverse geocoding technique is used to get the place information and save it to the photo of the user. The reverse geocoding technique is also used to find suggested keywords relevant to the place. Text-based image retrieval is used to retrieve the desired photo taken in a specific place.
View full abstract