Process variations in modern VLSI technologies are growing in both magnitude and dimensionality. To assess performance variability, complex simulation and performance models param...
Abstract. Nearest-neighbor queries in high-dimensional space are of high importance in various applications, especially in content-based indexing of multimedia data. For an optimiz...
Many classes of image data span a low dimensional nonlinear space embedded in the natural high dimensional image space. We adopt and generalize a recently proposed dimensionality ...
We provide evidence that non-linear dimensionality reduction, clustering and data set parameterization can be solved within one and the same framework. The main idea is to define ...
In solving complex visual learning tasks, adopting multiple descriptors to more precisely characterize the data has been a feasible way for improving performance. These representa...