TY - GEN
T1 - A system for landscape photograph localization
AU - Suzuki, Tetsuya
AU - Tokuda, Takehiro
PY - 2006/12/1
Y1 - 2006/12/1
N2 - We present a system which supports localization of land-scape photographs. Localization of a photograph is to assign the camera parameters to the photograph. The parameters are the camera's position, direction, focus and film size to take the photograph. An application of the system is metadata generation for location oriented search systems. Target photographs for the system are landscape photographs such that terrain in the photographs is enough characteristic to specify the locations. We describes features of photographic objects and a search area to define a search problem. The system takes the problem definition and an initial solution set as its input, and improves the solutions by a specified algorithm. We can use a genetic algorithm and a local search algorithm as the algorithm. Key points to the system are what we can describe to define search problems, and how effectively the system searches good solutions. Two evaluation experiments under an ideal condition shows that its expressiveness is enough and the system can find appropriate solutions effectively.
AB - We present a system which supports localization of land-scape photographs. Localization of a photograph is to assign the camera parameters to the photograph. The parameters are the camera's position, direction, focus and film size to take the photograph. An application of the system is metadata generation for location oriented search systems. Target photographs for the system are landscape photographs such that terrain in the photographs is enough characteristic to specify the locations. We describes features of photographic objects and a search area to define a search problem. The system takes the problem definition and an initial solution set as its input, and improves the solutions by a specified algorithm. We can use a genetic algorithm and a local search algorithm as the algorithm. Key points to the system are what we can describe to define search problems, and how effectively the system searches good solutions. Two evaluation experiments under an ideal condition shows that its expressiveness is enough and the system can find appropriate solutions effectively.
UR - http://www.scopus.com/inward/record.url?scp=34547514633&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=34547514633&partnerID=8YFLogxK
U2 - 10.1109/ISDA.2006.80
DO - 10.1109/ISDA.2006.80
M3 - Conference contribution
AN - SCOPUS:34547514633
SN - 0769525288
SN - 9780769525280
T3 - Proceedings - ISDA 2006: Sixth International Conference on Intelligent Systems Design and Applications
SP - 1080
EP - 1085
BT - Proceedings - ISDA 2006
T2 - ISDA 2006: Sixth International Conference on Intelligent Systems Design and Applications
Y2 - 16 October 2006 through 18 October 2006
ER -