A good training dataset, representative of the test images expected in a given application, is critical for ensuring good performance of a visual categorization system. Obtaining ...
Aniruddha Kembhavi, Behjat Siddiquie, Roland Miezi...
Human matching between different fields of view is a difficult problem in intelligent video surveillance; whereas fusing multiple features has become a strong tool to solve it. In...
Abstract--Flash memory is well-known for its inherent asymmetry: the flash-cell charge levels are easy to increase but are hard to decrease. In a general rewriting model, the store...
Anxiao Jiang, Michael Langberg, Moshe Schwartz, Je...
The design of physical interfaces has been constrained by the relative akinesis of the material world. Current advances in materials science promise to change this. In this paper, ...
Kernel techniques have long been used in SVM to handle linearly inseparable problems by transforming data to a high dimensional space, but training and testing large data sets is ...