Kenneth Barner, Professor and Chairman
Prof. Barner's research interests focus on signal and image processing as well as human-computer interaction. The processing of signals is fundamental to applications in a broad array of disciplines, from biology and medicine, to communications and imaging. Traditional signal processing methods employ linear techniques. However, real-world systems and signals often exhibit nonlinearities. Moreover, linear methods break down in the presence of harsh environments, such as those characterized by impulsive, or heavy tailed distributions. Thus current work focuses on developing, analyzing, and employing nonlinear signal processing methods that are robust to harsh environments and are able to exploit system nonlinearities. Traditional signal processing methods also rely on Nyquist approaches to sampling and processing signals. This approach breaks down, or is inefficient, in real world cases, particularly those with multispectral components. Moreover, traditional approaches fail to exploit the sparsity inherent in many problems. Thus current work also leads and capitalizes on the growing body of sparsity-based approaches to sampling, processing, and detecting signals of interest. Sparsity and nonlinearity have been exploited in multiple theoretical and application focused projects, including the processing of biomedical signals, such as EEG and ECG signals and tomographic images for compression, feature extraction, and enhancement. Imaging, including stereo imaging, has been utilized for facial recognition, gait recognition, and human computer interface strategies, including those on mobile platforms.