Unlocking the Urban Photographic Record Through 4D Scene Understanding and Modeling

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

There is a growing need for novel ways to access the exponentially growing archives of historical imagery. It is imperative to go beyond cataloging, indexing, and keyword driven databases, to a paradigm where the computer at least partially understands the content of images. Pushing the state of the art in scene understanding and 3D modeling will enable radical new ways to view and experience historical and/or temporally varying imagery. This research aims at building time-varying 3D models that can serve to pull together large collections of images pertaining to the appearance, evolution, and events surrounding one place or artifact over time, as exemplified by the "4D City" project: the completely automatic construction of a 4D database showing the evolution over time of a single city. In terms of intellectual merits, the project will advance the state of the art in (1) single-view image interpretation for matching scene structure across space and time; (2) model-based, probabilistic inference and stochastic grammars to encode high-level prior knowledge; and (3) introducing time as an additional object of inference. The project is expected to broadly impact society in the areas of urban planning, historical education, research, and preservation. In addition, the newly developed algorithms can be applied in other application areas such as forensics and intelligence gathering. Finally, the planned dissemination efforts will center on an instantiation of the 4D City project in Atlanta, and will touch many through web-based delivery and a kiosk installation in the Atlanta History Center. http://4d-cities.cc.gatech.edu