Real time convolution has many applications among others simulating room reverberation in audio processing. Non-uniformly partitioning filters could satisfy the both desired features of having a low latency and less computational complexity for an efficient convolution. However, distributing the computation to have an uniform demand on Central Processing Unit (CPU) is still challenging. Moreover, computational cost for very long filters is still not acceptable. In this thesis, a new algorithm is presented by taking advantage of the broad memory on Graphics Processing Units (GPU). Performing the computations of a non-uniformly partitioned block convolution on GPU could solve the problem of work load on CPU. It is shown that the computational time in this algorithm reduces for the filters with long length.

This thesis investigated a method of barcode reconstruction to address the recovery of a blurred and convoluted one-dimensional barcode. There are a lot of types of barcodes used today, such as Code 39, Code 93, Code 128, etc. Our algorithm applies to the universal barcode, EAN 13. We extend the methodologies proposed by Iwen et al. (2013) in the journal article "A Symbol-Based Algorithm for Decoding barcodes." The algorithm proposed in the paper requires a signal measured by a laser scanner as an input. The observed signal is modeled as a true signal corrupted by a Gaussian convolution, additional noises, and an unknown multiplier. The known barcode dictionaries were incorporated into the forward map between the true barcode and the observed barcode. Unlike the one proposed by Iwen et al., we take dictionaries of different patterns into account, specifically for decoding book barcodes from images which are captured with smartphones. We also presented numerical experiments that examined the performance of the proposed algorithm and illustrated that the unique determination of barcode digits is possible even in the presence of noise.

The method development process in liquid chromatography (LC) involves optimization of a variety of method parameters including stationary phase chemistry, column temperature, initial and final mobile phase compositions, and gradient time when gradient mobile phases are used. Here, a general simulation program to predict the results (i.e., retention time, peak width and peak shape) of LC separations, with the ability to study various complex chromatographic conditions is described. The simulation program is based on the Craig distribution model where the column is divided into discrete distance (Δz) and time (Δt) segments in a grid and is based on parameterization with either the linear solvent strength or Neue-Kuss models for chromatographic retention. This algorithm is relatively simple to understand and produces results that agree well with closed form theory when available. The set of simulation programs allows for the use of any eluent composition profile (linear and nonlinear), any column temperature, any stationary phase composition (constant or non-constant), and any composition and shape of the injected sample profile. The latter addition to our program is particularly useful in characterizing the solvent mismatch effect in comprehensive two-dimensional liquid chromatography (2D-LC), in which there is a mismatch between the first dimension (1D) effluent and second dimension (2D) initial mobile phase composition. This solvent mismatch causes peak distortion and broadening. The use of simulations can provide a better understanding of this phenomenon and a guide for the method development for 2D-LC. Another development that is proposed to have a great impact on the enhancement of 2D-LC methods is the use of continuous stationary phase gradients. When using rapid mobile phase gradients in the second dimension separation with diode array detection (DAD), refractive index changes cause large backgrounds such as an injection ridge (from solvent mismatch) and sloping baselines which can be problematic for achieving accurate quantitation. Use of a stationary phase gradient may enable the use of an isocratic mobile phase in the 2D, thus minimizing these background signals. Finally, our simulator can be used as an educational tool. Unlike commercially available simulators, our program can capture the evolution of the chromatogram in the form of movies and/or snapshots of the analyte distribution over time and/or distance to facilitate a better understanding of the separation process under complicated circumstances. We plan to make this simulation program publically available to all chromatographers and educators to aid in more efficient method development and chromatographic training.

This dissertation focuses on the development of three classes of brain-inspired machine learning classification models. The models attempt to emulate (a) multi-sensory integration, (b) context-integration, and (c) visual information processing in the brain.The multi-sensory integration models are aimed at enhancing object classification through the integration of semantically congruent unimodal stimuli. Two multimodal classification models are introduced: the feature integrating (FI) model and the decision integrating (DI) model. The FI model, inspired by multisensory integration in the subcortical superior colliculus, combines unimodal features which are subsequently classified by a multimodal classifier. The DI model, inspired by integration in primary cortical areas, classifies unimodal stimuli independently using unimodal classifiers and classifies the combined decisions using a multimodal classifier. The multimodal classifier models are be implemented using multilayer perceptrons and multivariate statistical classifiers. Experiments involving the classification of noisy and attenuated auditory and visual representations of ten digits are designed to demonstrate the properties of the multimodal classifiers and to compare the performances of multimodal and unimodal classifiers. The experimental results show that the multimodal classification systems exhibit an important aspect of the “inverse effectiveness principle” by yielding significantly higher classification accuracies when compared with those of the unimodal classifiers. Furthermore, the flexibility offered by the generalized models enables the simulations and evaluations of various combinations of multimodal stimuli and classifiers under varying uncertainty conditions. The context-integrating model emulates the brain’s ability to use contextual information to uniquely resolve the interpretation of ambiguous stimuli. A deep learning neural network classification model that emulates this ability by integrating weighted bidirectional context into the classification process is introduced. The model, referred to as the CINET, is implemented using a convolution neural network (CNN), which is shown to be ideal for combining target and context stimuli and for extracting coupled target-context features. The CINET parameters can be manipulated to simulate congruent and incongruent context environments and to manipulate target-context stimuli relationships. The formulation of the CINET is quite general; consequently, it is not restricted to stimuli in any particular sensory modality nor to the dimensionality of the stimuli. A broad range of experiments are designed to demonstrate the effectiveness of the CINET in resolving ambiguous visual stimuli and in improving the classification of non-ambiguous visual stimuli in various contextual environments. The fact that the performance improves through the inclusion of context can be exploited to design robust brain-inspired machine learning algorithms. It is interesting to note that the CINET is a classification model that is inspired by a combination of brain’s ability to integrate contextual information and the CNN, which is inspired by the hierarchical processing of visual information in the visual cortex. A convolution neural network (CNN) model, inspired by the hierarchical processing of visual information in the brain, is introduced to fuse information from an ensemble of multi-axial sensors in order to classify strikes such as boxing punches and taekwondo kicks in combat sports. Although CNNs are not an obvious choice for non-array data nor for signals with non-linear variations, it will be shown that CNN models can effectively classify multi-axial multi-sensor signals. Experiments involving the classification of three-axis accelerometer and three-axes gyroscope signals measuring boxing punches and taekwondo kicks showed that the performance of the fusion classifiers were significantly superior to the uni-axial classifiers. Interestingly, the classification accuracies of the CNN fusion classifiers were significantly higher than those of the DTW fusion classifiers. Through training with representative signals and the local feature extraction property, the CNNs tend to be invariant to the latency shifts and non-linear variations. Moreover, by increasing the number of network layers and the training set, the CNN classifiers offer the potential for even better performance as well as the ability to handle a larger number of classes. Finally, due to the generalized formulations, the classifier models can be easily adapted to classify multi-dimensional signals of multiple sensors in various other applications.

Visual object tracking is a classical and very popular problem in computer vision with a plethora of applications such as vehicle navigation, human computer interface, human motion analysis, surveillance, auto-control systems and many more. Given the initial state of a target in the first frame, the goal of tracking is to predict states of the target over time where the states describe a bounding box covering the target. Despite numerous object tracking methods that have been proposed in recent years [1-4], most of these trackers suffer a degradation in performance mainly because of several challenges that include illumination changes, motion blur, complex motion, out of plane rotation, and partial or full occlusion, while occlusion is usually the most contributing factor in degrading the majority of trackers, if not all of them. This thesis is devoted to the advancement of generic object trackers tackling different challenges through different proposed methods. The work presented propose four new state-of-the-art trackers. One of which is 3D based tracker in a particle filter framework where both synchronization and registration of RGB and depth streams are adjusted automatically, and three works in correlation filters that achieve state-of-the-art performance in terms of accuracy while maintaining reasonable speeds.

The Diploma thesis deals with object surface reconstruction by means of confocal microscope. It includes part of mathematical theory which is associated with this problem. The mathematical algorithm of the object surface reconstruction is illustrated. This Diploma thesis result is the application development for this reconstruction. The thesis also include the results of specific object.

Tato magisterská práce se zabývá využitím konvolučních neuronových sítí pro segmentační účely v oblasti transmisní elektronové mikroskopie. Také popisuje zvolenou topologii neuronové sítě - U-NET, použíté augmentační techniky a programové prostředí. Firma Thermo Fisher Scientific (dříve FEI Czech Republic s.r.o) poskytla obrazová data pro účely této práce. Získané segmentační výsledky jsou prezentovány ve formě křivek (ROC, PRC) a ve formě numerických hodnot (ARI, DSC, Chybová matice). Zvolená UNET topologie dosáhla excelentních výsledků v oblasti pixelové segmentace. S největší pravděpodobností, budou tyto výsledky sloužit jako odrazový můstek pro interní firemní výzkum.

The Diploma thesis deals with reconstruction of Missing data of an Image. It is done by the use of appropriate Mathematical theory and numerical algorithm to reconstruct missing information. The result of this implementation is the reconstruction of missing image information. The thesis also compares different numerical methods, and see which one of them perform best in terms of efficiency and accuracy of the given problem, hence it is used for the reconstruction of missing data.

Thesis provides an overview and discussion of current findings in the field of biometrics. In particular, it focuses on facial recognition subject. Special attention is payed to convolutional neural networks and capsule networks. Thesis then lists current approaches and state-of-the-art implementations. Based on these findings it provides insight into engineering a very own solution based of CapsNet architecture. Moreover, thesis discussed advantages and capabilitied of capsule neural networks for identification of a person by its face.

This diploma thesis deals with the detection of fire and smoke from the image signal. The approach of this work uses a combination of convolutional and recurrent neural network. Machine learning models created in this work contain inception modules and blocks of long short-term memory. The research part describes selected models of machine learning used in solving the problem of fire detection in static and dynamic image data. As part of the solution, a data set containing videos and still images used to train the designed neural networks was created. The results of this approach are evaluated in conclusion.