Graphic Object Retrieval from Remote Database for Mobile Devices : Application to Building Images
building recognition, edge detection, mobile
The goal of the project is to create an image based navigation and identification system. It
consists of a client mobile device (i.e. PDA or telephone with camera) where a user can take
a picture of an object and send this picture with or without additional data (i.e GPS
information, UMTS cell location) to a server.
The server analyzes the image of the object to extract its features. It then uses these features and the additional data to perform an image retrieval process. The extraction of image features and retrieval process are the main focus of this PhD project.
We aim to develop an automatic method (i.e. no user intervention) that is capable of
recognizing man-made object (i.e. building) in a relatively simple, fast and efficient way, as we intended for this method to be implemented in mobile devices. We have chosen to work on three image features primitives ; color, shape and texture.