Image based search in mobile touristic information systems (in cooperation with ECCA - etourism competence center austria)

Thesis Type Bachelor
Thesis Status
Finished
Student Daniel Winkler
Init
Final
Start
Thesis Supervisor
External Supervisor
W. Höpken
Contact

The idea of this thesis is to make content based image retrieval available for touristic issues. A term of use therefore would be a tourist, who identifies some building as sight. For example 50 other tourists standing before buildings and taking photos of buildings could give him hints to detect such sights.
But now the tourist has the problem that he found a sight, but he can't get information about this sight with his mobile device since he doesn't know the name of this sight. This comes from the circumstance that searching the internet is mostly based on text-retrieval, and therefore the tourist has just the possibility to search for information by entering the name or at least some keywords describing the sight.

To help this tourist the idea of searching information by content based image retrieval (CBIR) came up. This would allow the tourist to identify the sight by taking a photo of it. He'd just have take a photo with his mobile device, send an MMS and would get back an MMS with information for the sight seconds later.

Since the realization of such a system from scratch would be to much for a bachelor thesis, this thesis confines to the evaluation of existing systems and checking the feasibility of the just called idea.