This course is part of our eLearning Archive, which includes older courses that may not be current or as user-friendly as courses designed more recently. This course examines Information Theory and our efforts to develop an information-theoretic criterion which can be utilized in adaptive filtering and neurocomputing. The main aim of our research is to develop new signal processing techniques by going beyond the basic assumptions of Linearity, Gaussianity, and Stationarity. By capturing higher order statistics of data using Information Theory, we solve a variety of problems in Biomedical Signal Processing, Communications, and Machine Learning.
This tutorial introduces issues in long-range facial image acquisition and measures for image quality and their usage, as well as subsequent challenges for face recognition. The first several modules on image acquisition for face recognition discuss concerns related to lighting, sensors, and lenses, which impact short-range biometrics, but are more pronounced in long-range biometrics. We then go on to introduce the design of controlled experiments for long-range face recognition and why they are needed. With our experiments, we go on to show some of the weather and atmospheric effects that occur for long-range imaging, with numerous examples. Next, we address measurements of "system quality" including image-quality measures and their use in the prediction of face recognition algorithm performance. That module introduces the concept of post-recognition score analysis and techniques for analyzing different "equality" measures. The last two modules of this tutorial explore long-range face recognition directly. Facial feature detection is an important prerequisite for face recognition, and we look at two different approaches for accurately accomplishing this for long-range scenarios. Finally, we address the very difficult problem of blur "both motion and atmospheric" including common sources in acquisition and algorithms to mitigate its effects.
In this tutorial, we're going to cover the background to gait as a biometric. We're going to consider studies in medicine, psychology, and literature, all of which have evidence that people can be recognized by the way that they walk. We are then going to look at databases that we can use to evaluate gait recognition potential. We will cover the three main databases that are in current use. The main techniques will be covered. We'll look at silhouette based techniques and model based techniques. These are the two main approaches, as they are in many biometrics. Although we have model-based face and holistic face, in gait recognition we have silhouette based and model based. We're then going to consider extensions to these techniques. Among the most important is viewpoint invariance. We will look at normalizing the signatures to be invariant to the relative viewpoint of the camera to the subject. We are then going to see how different shoes and different clothes affect the way you walk. In addition, we will look at a new database and a new system. We are going to cover the main components in automatic gait recognition, the main techniques, and the requirements of this new technology. We will use many sources of information, and so there are apologies if your own is missed, or perhaps your favorite technique is not covered. The references will either be on the page in which they are used, or the main ones will be collected together at the end of this talk. The aim is to show you that gait is a large and rapidly advancing field of research.
This course covers the rapidly evolving field of Computational Intelligence and focuses on the current understanding of the fundamental principles of working the mind, their computational implementations, and practical applications. This tutorial covers mind mechanisms including concepts, emotions, instincts, behavior, language, cognition, understanding, thinking, intuitions, conscious and unconscious abilities for formation of symbols and aesthetic feelings. Computational techniques are given for these mechanisms and abilities. The goal of this tutorial is to provide a basic mathematical understanding of the working of the mind. Its secondary goal is to demonstrate practical applications of these mechanisms for pattern recognition, tracking, fusion, search engines, and for integrated systems combining sensor signals and communication data. Lastly, this tutorial will outline future research directions. Historical and current difficulties in developing intelligent systems (IS) and applications will be discussed along with how the mind and new computational techniques overcome these difficulties.