At the Skoghall pappermill, sulphatepaper pulp is produced in a continuous digester originally from 1969. To be able to maintain a high level of production there is a need for a process with few disturbances. Variations in how well the wooden chips are packed in the digester is one form of disturbance. Today there are no available measurements on how well the chips are packed. Instead this is regarded as being constant. The variation in the so called bulk density of the chips is mainly due to variations in the percentage with small dimensions. Chips are classified in relation to their size and one of the smallest classes is referred to as pin chips. These are believed to have a big impact on the bulk density. The amount of pin chips fluctuate more then the other classes, there by causing disturbances. The Skoghall pappermill has invested in a ScanChip. This is an instrument that measures the dimensions of the chips optically. ScanChip presents figures on chip quality, including a measurement of the bulk density. However, it has been shown that this measurement is not valid for the Skoghall pappermill. By using data from ScanChip a model that predicts how well the chips are packed has been devised. This value is the bulk density divided by the basic density. The model has proved to yield good results, despite a relatively small amount of data. A theoretical value of the amount of produced pulp has been computed based on the revolutions of the production screw that feeds chips into the digester. This value takes in consideration how well the chips are packed. The value has shown great similarities with the empirical measurements that are used today. A simulation during one month has shown that differences in the mixture of chips have effected the measurement of produced pulp with up to 7 ton/h. Chips are stored in open pile storages before they are being used in the process of transforming them into pulp. Four screws are used to move chips from the piles to conveyer belts. It has been shown in work done previously, that the movement of the screws contributes to variations in the amount of pin chips measured by ScanChip. During the work with this master’s thesis I have found that there are variations in the piles that make it difficult to predict the amount of pin chips accordingly. However by filtering the measurements of pin chips to remove these variations, the results are improved. A new way of controlling the movements of the screws was operational on the 10 of March and this improved the results. The direction in which the screws are moving influence the speed of the screws, mainly in the pile with the so called sawmill chips. By changing the amount of chips that each screw puts out, the differences in speed have been reduced. The mixtures found in the two piles are not completely homogenous. There are a greater amount of pin chips in the northern parts compared with the southern parts. This could be an effect of the wind direction, and will still cause variations.

A Novelty Detection Approach to Seizure Analysis from Intracranial EEG Andrew B. Gardner 146 pages Directed by Dr. George Vachtsevanos and Dr. Brian Litt A framework for support vector machine classification of time series events is proposed and applied to analyze physiological signals recorded from epileptic patients. In contrast to previous works, this research formulates seizure analysis as a novelty detection problem which allows seizure detection and prediction to be treated uniformly, in a way that is capable of accommodating multichannel and/or multimodal measurements. Theoretical properties of the support vector machine algorithm employed provide a straightforward means for controlling the false alarm rate of the detector. The resulting novelty detection system was evaluated both offline and online on a corpus of 1077 hours of intracranial electroencephalogram (IEEG) recordings from 12 patients diagnosed with medically resistant temporal lobe epilepsy during evaluation for epilepsy surgery. These patients collectively had 118 seizures during the recording period. The performance of the novelty detection framework was assessed with an emphasis on four key metrics: (1) sensitivity (probability of correct detection), (2) mean detection latency, (3) early-detection fraction (prediction or detection of seizure prior to electrographic onset), and (4) false positive rate. Both the offline and online novelty detectors achieved state-of-the-art seizure detection performance. In particular, the online detector achieved 97.85% sensitivity, -13.3 second latency, and 40% early-detection fraction at an average of 1.74 false positive predictions per hour (Fph). These results demonstrate that a novelty detection approach is not only feasible for seizure analysis, but it improves upon the state-of-the-art as an effective, robust technique. Additionally, an extension of the basic novelty detection framework demonstrated its use as a simple, effective tool for examining the spread of seizure onsets. This may be useful for automatically identifying seizure focus channels in patients with focal epilepsies. It is anticipated that this research will aid in localizing seizure onsets, and provide more efficient algorithms for use in a real device.

In this thesis, a popular music genre recognition approach for Japanese popular music using SVM (support vector machine) with forward feature selection is proposed. First, various common acoustic features are extracted from the digital signal of popular music songs, including sub-bands, energy, rhythm, tempo, formants. A set of the most appropriate features for the genre identification is then selected by the proposed forward feature selection technique. Experiments conducted on the database consisting of 296 Japanese popular music songs demonstrate that the accuracy of recognition the proposed algorithm can achieve approximately 78.81% and the accuracy is stable when the number of testing music songs is increased.

In this thesis, we present a vehicle detection and tracking system based on image processing and pattern recognition of single camera image sequences. Both software design and hardware implementation are considered. In the hypothesis generation (HG) step and the hypothesis verification (HV) step, we use the shadow detection technique combined with the proposed constrained vehicle width/distance ratio to eliminate unreasonable hypotheses. Furthermore, we use SVM classifier, a popular machine learning technique, to verify the generated hypothesis more precisely. In the vehicle tracking step, we limit vehicle tracking duration and periodic vehicle detection mechanisms. These tracking methods alleviate our driver-assistant system from executing complex operations of vehicle detection repeatedly and thus increase system performance without sacrificing too much in case of tracking wrong objects. Based on the the profiling of the software execution time, we implement by hardware the most critical part, the preprocessing of intensity conversion and edge detection. The complete software/hardware embedded system is realized in a FPGA prototype board, so that performance of whole system could achieve real-time processing without too much hardware cost.

In this thesis, least squares support vector machine for regression (LS-SVR) is applied to image coding. First, we propose five simple algorithms for solving LS-SVR. For linear regression, two simple Widrow-Hoff-like algorithms, in primal form and in dual form, are proposed for LS-SVR problems. The dual form of the algorithm is then generalized to kernel-based nonlinear LS-SVR. The elegant and powerful two-parameter sequential minimization optimization (2PSMO) and three-parameter sequential minimization optimization (3PSMO) algorithms are provided in detail. A predictive function obtained from LS-SVR is utilized to approximate the gray levels of the image. After pruning, only a subset of training data called support vectors is saved. Experimental results on seven image blocks show that the LS-SVR with Gaussian kernel is more appropriate than that with Mahalanobis kernel with a covariance matrix. Two-layer LS-SVR is proposed to choose the machine parameters of the LS-SVR. Before training outer LS-SVR, feature extraction is used to reduce the input dimensionality. Experimental results on three whole images show that the results with two-layer LS-SVR after reducing dimensionality are better than those with two-layer LS-SVR without reducing dimensionality in PSNR for Lena and Baboon images and they are almost the same in PSNR for F16 image.

In this study, a 3D face representation and recognition method based on spherical harmonics expansion is proposed. The input data to the method is range image of the face. This data is called 2.5 dimensional. Input faces are manually marked on the two eyes, nose and chin points. In two dimensions, using the marker points, the human face is modeled as two concentric half ellipses for the selection of region of interest. These marker points are also used in three dimensions to register the faces so that the nose point tip is at the origin and the line across the two eyes lies parallel to the horizontal plane. A PCA based component analysis is done to further align the faces vertically. The aligned face is stitched and mapped to an ellipsoid and transformed using real spherical harmonics expansion. The real harmonics expansion coefficients are labeled and stored into a gallery. Using these coefficients as input, several classification algorithms are applied and the results are reported.

One of the major problems encountered in wireless location is the effect caused by non-line of sight (NLOS) propagation. When the direct path from the mobile station (MS) to base stations (BSs) is blocked by obstacles or buildings, the signal arrival times will delay. That will make the signal measurements include an error due to the excess path propagation. If we use the NLOS signal measurements for localization, that will make the system localization performance reduce greatly. In the thesis, a time-of-arrival (TOA) based location system with NLOS mitigation algorithm is proposed. The proposed method uses least squares-support vector machine (LS-SVM) with optimal parameters selection by particle swarm optimization (PSO) for establishing regression model, which is used in the estimation of propagation distances and reduction of the NLOS propagation errors. By using a weighted objective function, the estimation results of the distances are combined with suitable weight factors, which are derived from the differences between the estimated measurements and the measured measurements. By applying the optimality of the weighted objection function, the method is capable of mitigating the NLOS effects and reducing the propagation range errors. Computer simulation results in ultra-wideband (UWB) environments show that the proposed NLOS mitigation algorithm can reduce the mean and variance of the NLOS measurements efficiently. The proposed method outperforms other methods in improving localization accuracy under different NLOS conditions.

La multiplication des conflits de part le monde a eu pour principale conséquence de disséminer des millions de mines antipersonnel qui mettent en danger la vie des populations et constituent une entrave au développement économique des régions concernées. Dans ce mémoire, de nouvelles techniques de traitement du signal sont proposées pour la détection des mines antipersonnel dans les données enregistrées par un GPR (Ground Penetrating Radar). Deux filtres numériques sont tout d'abord conçus pour réduire le clutter, qui constitue un ensemble de phénomènes indésirables, dans les données de type Bscan et Cscan fournies par le GPR. Ces deux types de données représentent respectivement des images de tranches verticales et horizontales du sous-sol. La conception des filtres nécessite une modélisation géométrique du clutter et d'une signature de mine, pour chaque type de données, suivie d'une analyse spectrale permettant de définir le gabarit du filtre idéal. Ensuite, une nouvelle méthode de détection, basée sur une technique de détection de ruptures non paramétrique, est proposée afin de localiser automatiquement les réponses des mines antipersonnel sur des données Bscan. La méthode consiste à rechercher les ruptures spatiales suivant la direction des mesures afin de détecter les positions horizontales des mines et les ruptures suivant l'axe temporel afin de détecter les temps de réponse des mines. Une méthode de détection, basée sur l'extraction de contours fermés, est également proposée pour localiser les réponses des mines sur des données Cscan. Les performances de ces deux méthodes de détection sont évaluées par le calcul de probabilité de détection et de fausses alarmes.

Gene finding in eukaryotic genomes is an essential part of a comprehensive approach to modern systems biology. Most methods developed in the past rely on a combination of computational prediction and external information about gene structures from transcript sequences and comparative genomics. In the past, external sequence information consisted of a combination of full-length cDNA and expressed sequence tag (EST) sequences. Much improvement in prediction of genes and gene isoforms is promised by availability of RNA-seq data. However, productive use of RNA-seq for gene prediction has been difficult due to challenges associated with mapping RNA-seq reads which span splice junctions to prevalent splicing noise in the cell. This work addresses this difficulty with the development of methods and implementation of two new pipelines: 1/ a novel pipeline for accurate mapping of RNA-seq reads to compact genomes and 2/ a pipeline for prediction of genes using the RNA-seq spliced alignments in eukaryotic genomes. Machine learning methods are employed in order to overcome errors associated with the process of mapping short RNA-seq reads across introns and using them for determining sequence model parameters for gene prediction. In addition to the development of these new methods, genome annotation work was performed on several plant genome projects.

This thesis examines the application of a Support Vector Machine (SVM) classifier to automatically detect sleep and quiet wake (rest) behavior in mice from pressure signals on their cage floor. Previous work employed Neural Networks (NN) and Linear Discriminant Analysis (LDA) to successfully detect sleep and wake behaviors in mice. Although the LDA was successful in distinguishing between the sleep and wake behaviors, it has several limitations, which include the need to select a threshold and difficulty separating additional behaviors with subtle differences, such as sleep and rest. The SVM has advantages in that it offers greater degrees of freedom than the LDA for working with complex data sets. In addition, the SVM has direct methods to limit overfitting for the training sets (unlike the NN method). This thesis develops an SVM classifier to characterize the linearly non separable sleep and rest behaviors using a variety of features extracted from the power spectrum, autocorrelation function, and generalized spectrum (autocorrelation of complex spectrum). A genetic algorithm (GA) optimizes the SVM parameters and determines a combination of 5 best features. Experimental results from over 9 hours of data scored by human observation indicate 75% classification accuracy for SVM compared to 68% accuracy for LDA.