Beate Meffert

EEG artifact elimination by extraction of ICA-component features using image processing algorithms.

Artifact rejection is a central issue when dealing with electroencephalogram recordings. Although independent component analysis (ICA) separates data in linearly independent components (IC), the classification of these components as artifact or EEG signal still requires visual inspection by experts. In this paper, we achieve automated artifact elimination using linear discriminant analysis (LDA) for classification of feature vectors extracted from ICA components via image processing algorithms. We compare the performance of this automated classifier to visual classification by experts and identify range filtering as a feature extraction method with great potential for automated IC artifact recognition (accuracy rate 88%). We obtain almost the same level of recognition performance for geometric features and local binary pattern (LBP) features. Compared to the existing automated solutions the proposed method has two main advantages: First, it does not depend on direct recording of artifact signals, which then, e.g. have to be subtracted from the contaminated EEG. Second, it is not limited to a specific number or type of artifact. In summary, the present method is an automatic, reliable, real-time capable and practical tool that reduces the time intensive manual selection of ICs for artifact removal. The results are very promising despite the relatively small channel resolution of 25 electrodes.

Infrared-monitored cold response in the assessment of Raynaud's phenomenon.

Background.  Evaluation of treatments for Raynauds phenomenon (RP) requires objective response parameters in addition to clinical activity scores. Thermographic monitoring of fingertip re‐warming after cold challenge has been widely used but usually requires sophisticated equipment. We have previously shown that fingertip re‐warming after cold challenge follows a first‐order transient response curve that can be described by a single variable, designated τ.

Towards Real Life Applications in Emotion Recognition

In this paper different kinds of emotional speech corpora are compared in terms of speech acquisition (acted speech vs. elicited speech), utterance length and similarity to spontaneous speech. Feature selection is applied to find an optimal feature set and to examine the correlation of different kinds of features to dimensions in the emotional space. The influence of different feature sets is evaluated. To cope with environmental conditions and to get a robust application, effects related to energy and additive noise are analyzed.

Automated EEG artifact elimination by applying machine learning algorithms to ICA-based features

OBJECTIVE Biological and non-biological artifacts cause severe problems when dealing with electroencephalogram (EEG) recordings. Independent component analysis (ICA) is a widely used method for eliminating various artifacts from recordings. However, evaluating and classifying the calculated independent components (IC) as artifact or EEG is not fully automated at present. APPROACH In this study, we propose a new approach for automated artifact elimination, which applies machine learning algorithms to ICA-based features. MAIN RESULTS We compared the performance of our classifiers with the visual classification results given by experts. The best result with an accuracy rate of 95% was achieved using features obtained by range filtering of the topoplots and IC power spectra combined with an artificial neural network. SIGNIFICANCE Compared with the existing automated solutions, our proposed method is not limited to specific types of artifacts, electrode configurations, or number of EEG channels. The main advantages of the proposed method is that it provides an automatic, reliable, real-time capable, and practical tool, which avoids the need for the time-consuming manual selection of ICs during artifact removal.

FPGA-based Smart Camera for 3D wavelet-based image segmentation

In recent years hardware implementation of the wavelet transform for image processing has gathered significant attention. The wavelet transform is a promising tool for image processing and in particular for image compression and image segmentation. In this paper the 3D wavelet transform is used to detect moving objects or object groups as regions of interest (ROI). We propose a hardware implementation of the 2D wavelet transform, including PAL/NTSC video grabbing and Ethernet interface. It requires a low density FPGA board like V2PRO, while the 3D wavelet transform needs slightly more internal memory. Careful selection of those subbands which contain important coefficients reduces the computing power and memory requirements for video segmentation and movement detection. In addition, using the ROI mask the network transfer rate can be reduced significantly. The design is called smart camera because it delivers processed images instead of original ones.

Resolution mosaic-based Smart Camera for video surveillance

Video surveillance is one of the most data intensive applications. A typical video surveillance system consists of one or multiple video cameras, a central storage unit, and a central processing unit. At least two bottlenecks exist: First, the transmission capacity is limited, especially for raw data. Second, the central processing unit has to process the incoming data to give results in real time. Therefore, we propose an FPGA-based embedded camera system which performs all steps of image acquisition, region of interest extraction, generation of a multiresolution image, and image transmission. The proposed pipeline-based architecture allows a real time wavelet-based image segmentation and a detection of moving objects for surveillance purposes. The system is integrated in a single FPGA using external RAM as storage for images and for a Linux operating system which controls the data flow. With the pipeline concept and a Linux device driver it is possible to create a system for streaming the results of an image processing through a GbE interface. A real time processing is achieved. The camera transmits the captured images with 30 Mpixel/s, but the system is able to process 100 Mpixel/s.

Daubechies Versus Biorthogonal Wavelets for Moving Object Detection in Traffic Monitoring Systems

Moving object detection is a fundamental task for a variety of traffic applications. In this paper the Daubechies and biorthogonal wavelet families are exploited for extracting the relevant movement information in moving image sequences in a 3D wavelet-based segmentation algorithm. The proposed algorithm is applied for traffic monitoring systems. The objective and subjective experimental results obtained by applying both wavelet types are compared and interpreted in terms of the different wavelet properties and the characteristics of the image sequences. The comparisons show the superior performance of the symmetric biorthogonal wavelets in the presence of noisy images and changing lighting conditions when compared to the application of high order Daubechies wavelets. The algorithm is evaluated using simulated images in the Matlab environment.

Modified Maxwell Equations in Quantum Electrodynamics

Monopole, dipole, and multipole currents Hamiltonian formalism quantization of the pure radiation field Klein-Gordon equation and vacuum constants.

Evaluation based combining of classifiers for monitoring honeybees

We report on an image classification task originated from the video observation of beehives. Biologists desire to have an automatic support to identify individual bees which are labelled with badges. Current state of the art in object detection and evaluation of classifiers is briefly reviewed. Different algorithms are evaluated. ROC- as well as precision-recall analysis show that a gradient based method performs best. We investigate, whether and how this superior method can be further improved. Therefore, a novel approach for combining classifiers based on an evaluation methodology is proposed. From the pool of classifiers those are selected which provide complementary information while operating with high precision and low recall. This approach shows superior performance compared to the stand-alone detectors. The suggested combination strategy is compared to other combining rules as well as to the oracle classifier.

Resolution mosaic EM algorithm for medical image segmentation

Multiresolution analysis is an established part of the human vision system. It builds different representations of an image with a spatial resolution adapted to the size of objects of interest and to its level of relevance. Multiresolution analysis is an efficient tool for image segmentation. It allows the processing of global features as well as local features in a corresponding proper scale. Furthermore, it simplifies the segmentation process, accelerates the computation, and improves the results. In this work we propose a segmentation algorithm that is based on the multiresolution analysis for the magnetic resonance images of the human brain. It has been evaluated against known algorithm from the literature. The image subject to segmentation is preprocessed to be represented in a mosaic of different resolutions, based on the distribution of the information contained in the image. Then the conventional EM algorithm is applied for the segmentation.