Dimensionality reduction techniques seek to represent a set of images as a set of points in a low dimensional space. Here we explore a video representation that considers a video ...
Abstract The notorious "dimensionality curse" is a wellknown phenomenon for any multi-dimensional indexes attempting to scale up to high dimensions. One well-known approa...
Non-linear dimensionality reductionmethods are powerful techniques to deal with
high-dimensional datasets. However, they often are susceptible to local minima
and perform poorly ...
Andreas Geiger (Karlsruhe Institute of Technology)...
We investigate reducing the dimensionality of image sets by using principal component analysis on wavelet coefficients to maximize edge energy in the reduced dimension images. Lar...