Visual Surveillance systems have greatly increased in past few years. Several methods have been proposed in order to improve the efficiency of Face Detection but still remains a challenging task due to various illumination, poses and occlusion conditions. In this paper, we propose a novel method for Face Detection where a decision boundary is defined for skin classifier based on training datase...

Evaluating curves on surfaces is a frequently occurring operation in a number of applications involving surface interrogations For example evaluating the intersection curve of two surfaces is critical to boundary B rep computation and in applications involving visibility and rendering the ability to evaluate the silhouette curve of surfaces is important While dealing with high degree surfaces t...

This paper presents an automatic registration system for aligning combined rangeintensity scan pairs. The overall approach is designed to handle several challenges including extensive structural changes, large viewpoint differences, repetitive structure, illumination differences, and flat regions. The technique is split into three stages: initialization, refinement, and verification. During ini...

Support Vector Domain Description (SVDD) is inspired by the Support Vector Classifier. It obtains a sphere shaped decision boundary with minimal volume around a dataset. This data description can be used for novelty or outlier detection. Our approach is always to minimize the volume of the sphere describing the dataset, while at the same time maximize the separability between the spheres. To bu...

Even with the advent of more sophisticated, data-hungry methods, boosted decision trees remain extraordinarily successful for fast rigid object detection, achieving top accuracy on numerous datasets. While effective, most boosted detectors use decision trees with orthogonal (single feature) splits, and the topology of the resulting decision boundary may not be well matched to the natural topolo...

The first term is the estimation error and measures the performance of the discrimination rule with respect to the best hypothesis in H. In previous lectures, we studied performance guarantees for this quantity when ĥn is ERM. Now, we also consider the approximation error, which captures how well a class of hypotheses {Hk} approximates the Bayes decision boundary. For example, consider the hist...

We describe our shot-boundary detection experiments for the TREC-10 video track, using a multitimescale shot-change detection algorithm.

A key step in the application of novelty detection techniques to high-dimensional data is the exploration of relationships that are typically not obvious when dimensions are examined independently. Visualisation techniques allow such exploration of the structure of the data set by mapping high-dimensional data into lower dimensionality for inspection. This paper discusses the application of the...

We present a new algorithm, the Time Dependent Phase Space Filter (TDPSF) which is used to solve time dependent Nonlinear Schrodinger Equations (NLS). The algorithm consists of solving the NLS on a box with periodic boundary conditions (by any algorithm). Periodically in time, we decompose the solution into a family of coherent states. Those coherent states which are outgoing are deleted, while...

In the inspection of treated water samples under microscope, knowing the average number of parasite (oo)cysts like Giardia and Cryptosporidium that exist in the samples is crucial as it tells whether the water is safe for consumption. Here, we introduce a new approach using a bidirectional contour tracing technique to segment and enumerate overlapping Cryptosporidium and Giardia (oo)cysts in mi...