Featured Projects
City Scale Geo-spatial Trajectory Estimation of a Moving Camera
In this work, we address the problem of extracting the geospatial trajectory of the camera from unconstrained videos recorded in a city. The proposed method is intended for user-uploaded videos available on the web, which typically include undesired recording defects, such as blurred or uninformative frames, abrupt changes in camera motion, zooming effect, dynamical environment of the scene such as vehicles and pedestrians occluding distinct features, and lack of information of the initial position and pose (e.g. meta data) where the video was recorded.
Reference:
Gonzalo Vaca-Castano, Amir Roshan Zamir and Mubarak Shah
"City Scale Geo-spatial Trajectory Estimation of a Moving Camera", CVPR 2012.
Amir Roshan Zamir and Mubarak Shah
"Accurate Image Localization Based on Google Maps Street View", ECCV 2010.
Motion Analysis on CLIF Dataset
The problem that this project deals with is that of detection and tracking of large number of objects in Wide Area Surveillance (or WAS) data. WAS is aerial data that is characterized by a very large field of view. The particular dataset that we used for this project is the CLIF dataset. CLIF stands for Columbus Large Image Format, the dataset is available here.
Reference:
Vladimir Reilly, Haroon Idrees and Mubarak Shah
"Detection and Tracking of Large Number of Targets in Wide Area Surveillance", ECCV 2010
Modeling Traffic Patterns for Vision based Surveillance Applications
With the proliferation of wide area video sensor networks, video surveillance, especially in public areas, is gaining importance at an unprecedented rate. We argue that over the period of its operation, an intelligent tracking system should be able to learn the scene from its observables and be able to improve its performance based on this model. The high level knowledge necessary to make such inferences derives from domain knowledge, past experiences, as well as scene geometry, learned traffic and target behavior patterns in the area, etc. This argument forms the basis of this project, where we model and learn the scene activity, observed by a static camera. The motion patterns of the objects in the scene are modeled as a multivariate non-parametric probability density function of spatio-temporal variables. Kernel density estimation is used to learn this model in a completely unsupervised fashion, by observing the trajectories of objects over extended periods of time.
Reference:
Imran Saleemi, Khurram Shafique and Mubarak Shah
"Probabilistic Modeling of Scence Dynamics for Applications in Visual Surveillance", IEEE TPAMI 2008.