UVicSpace : Browsing by Supervisor "Agathoklis, Panajotis"

Browsing by Supervisor "Agathoklis, Panajotis"

Now showing 1 - 20 of 32

1-D and 2-D digital filters design using model reduction and optimization methods for broadband beamforming and interference rejection
(2025) Omar, Abdussalam; Agathoklis, Panajotis; Shpak, Dale John
This thesis presents several design algorithms for nearly linear-phase one-dimensional (1-D) and two-dimensional (2-D) infinite impulse response (IIR) digital filters. Optimization techniques as well as model order reduction (MOR) filter design methods are considered in this study. For 1-D, finite impulse response (FIR) filters can achieve perfectly linear phase which makes them important in applications such as the field of audio signal processing where a flat delay characteristic may be desired. However, in most applications a perfectly linear phase response is not required and filters that have nearly linear phase response are quite acceptable. In such cases, IIR filters are more attractive than FIR filters. The design of IIR filters is more challenging than that of FIR filters because it results in a highly nonlinear objective function that requires sophisticated optimization methods. The 1-D optimization method proposed here solves the problem of approximating specified magnitude and linear-phase responses simultaneously. Since IIR filters can be designed to have nearly linear phase response in the passband, their passband group delay is usually considerably smaller than the delay of linear-phase FIR filters with equivalent magnitude responses. Meeting a required minimum stopband attenuation or a minimum deviation from the desired magnitude and phase responses in the passbands are common design constraints that can be handled by the proposed optimization method for 1-D IIR filter. Also, an important constraint in the design of IIR filters is the prescription of a maximum pole radius, which allows to guarantee the stability margin and low coefficient selectivity for the obtained filter for finite-precision implementations. These design specifications are consistent with the constraints which often arise in practical filter design problems. In this research work, an optimization method for solving this constrained 1-D IIR design problem is presented. The above optimization method used for designing 1-D IIR filters is extended to 2-D separable-denominator IIR digital filters with nearly linear phase in the passband. During the development of the proposed design techniques for 2-D digital filters, a special emphasis has been placed on their computational efficiency and a method for the design of 2-D IIR digital filters based on a balanced realization (BR) model order reduction technique is proposed. In this method, the initial design is a linear phase 2-D FIR filter realized in a 2-D state space model, which leads to a stable 2-D separable-denominator IIR filter with nearly linear phase in the passband. The model reduction method is based on structured controllability Ps and structured observability Qs gramians. These gramians are block-diagonal positive-definite matrices satisfying 2-D Lyapunov inequalities. An efficient general algorithm is developed to compute these matrices by minimizing the trace of Ps and the trace of Qs under Linear Matrix Inequalities (LMI) constraints. The use of these gramians ensures that the resulting 2-D IIR filter is a 2-D stable filter. Furthermore, the obtained nearly linear-phase 2-D IIR filter is more economical and computationally more efficient than the original 2-D FIR filter. Numerical examples using MATLAB show that the proposed method provides a good compromise between the filter selectivity and computational complexity when compared to existing techniques, making the results of this dissertation directly applicable to many practical applications. For example, in the field of array signal processing, 2-D digital filters having a fan-shaped filter in the passband emerge as powerful tools, particularly when employed as beamformers in scenarios where the Direction of Arrivals (DOAs) of the desired broadband Plane Waves (PWs) are known. In such cases, the designed 2-D FIR and IIR filters having a fan-shaped filter passband in the 2-D frequency domains are used as beamformers. Benefiting from the knowledge of DOAs of the desired broadband PWs, these filters are used to extract the signal of interest (SOI), suppress the interference, and reduce the noise corrupting the SOI. The successful implementation of 2-D FIR and IIR fan filters as beamformers not only enhances the rejection of the interference but also demonstrates its capability to reduce the effect of AWGN. This dual functionality holds significant implications for practical applications in digital signal processing, in which robustness against interference and noise is important. Simulation results demonstrate a good performance of the proposed beamformers and confirmed that the filters obtained using the proposed methods are capable of extracting and enhancing the desired 2-D broadband signals according to their directions of arrival under severe interference and noise.
A 3-D visualization system for serial microscope images
(2018-07-18) Li, Jianping; Agathoklis, Panajotis
A three-dimensional (3-D) visualization system for serial microscope images is developed with special reference to its application in microscopy and cell biology. The 3-D visualization system involves three process stages, namely data acquisition, volume data modeling and object rendering. The data acquisition part deals with collecting serial microscope images and is carried out by optical sectioning which records serial microscope images from the top to the bottom of a specimen. A new algorithm is proposed to computationally remove out-of-focus information from each recorded image of a specimen. This algorithm processes serial images independently and thus avoids computationally expensive 3-D convolution and 3-D Fourier transforms. Further, an extensive study of imaging properties of defocused microscopes is carried out in the thesis. The defocused point spread functions and optical transfer functions of transmitted light microscopes have been analyzed. An extensive comparison of the two approaches of obtaining these functions, namely direct measurements and theoretical calculations. is conducted. An interesting observation is made that the results of these two approaches correspond well with each other only for low magnification and low numerical aperture objective lenses. Modeling of serial microscope images is carried out by an isosurface modeling algorithm. This algorithm is based on the marching-cube algorithm with two modifications proposed in the thesis. One modification is to detect and prevent the redundancy existing in the original marching-cube algorithm. The other modification is to use the middle-point algorithm to avoid linear interpolation to locate the vertex of a polygon. Results show that the modified marching-cube algorithm proposed in the thesis significantly reduces the number of polygons generated and thus greatly increases the computation efficiency for surface generation and object rendering. Object rendering, which is to generate a realistic image, is carried out by using a C library SImple Polygon Processor (SIPP). A graphic user-interface is designed and implemented to facilitate the modeling and rendering processes. It offers various user-friendly functions, such as rotation, zooming and cutting, for close examination of the objects under study and can be used for visualizing various volume data. The interactive system together with the data acquisition algorithm form a 3-D visualization system for serial microscope images. The results of visualization show that the 3-D visualization system developed in this thesis realistically and efficiently reconstructs objects of interest from serial microscope images as well as from various volume data such as Computer Tomography (CT) and Magnetic Resonance Imaging (MRI) medical images.
3D space-time digital filtering for radio astronomy
(2010-06-14T20:54:18Z) Liyanage, Najith; Agathoklis, Panajotis; Bruton, Leonard T.
The development of large dense arrays of smaller antennas and moderate-sized antennas equipped with multi-beam arrays in the next generation of radio telescopes, such as the square kilometre array (SKA), is making 3D space-time (ST) filtering increasingly relevant for applications in radio astronomy. The next generation of radio telescopes, such as the SKA, are expected to achieve unprecedented levels of sensitivity, resolution, survey speed and field of view (FOV). In this regard, the mitigation of radio frequency interference (RFI) and the suppression of mutually coupled (MC) signals on dense antenna arrays are being identified as most challenging. The main contribution of the work presented in this thesis is the proposal of a 3D ST filtering approach to enhance signals of interest (SOIs) by attenuating over-the-horizon RFIs and MC signals that exist on dense phased arrays. For this purpose, a novel 3D linear phase filter bank structure consisting of I D FIR filters and 2D circularly symmetric FIR filters is proposed. The advantage of the proposed ST filtering approach is its inherent capability of broadband (BB) processing. The proposed 3D ST filtering approach is investigated on synthesised BB SOI, BB RFI and BB MC signals on dense aperture arrays (AAs) and focal plane arrays (FPAs). The results indicate successful mitigation of over-the-horizon BB RFIs and moderate suppression of BB MC signals, without significantly distorting the BB SOIs.
Advanced techniques for improving radar performance
(2019-12-03) Shoukry, Mohammed Adel; Agathoklis, Panajotis
Wideband beamforming have been widely used in modern radar systems. One of the powerful wideband beamforming techniques that is capable of achieving a high selectivity over a wide bandwidth is the nested array (NA) beamformer. Such a beamformer consists of nested antenna arrays, 2-D spatio-temporal filters, and multirate filterbanks. Speed of operation is bounded by the speed of the hardware implementation. This dissertation presents the use of a systematic methodology for design space exploration of the NA beamformer basic building blocks. The efficient systolic array design in terms of the highest possible clock speed of each block was selected for hardware implementation. The proposed systolic array designs and the conventional designs were implemented in FPGA hardware to verify their functionality and compare their erformance. The implementations results confirm that the proposed systolic array implementations are faster and requires less hardware resources than the published designs. The overall beamformer FPGA implementation is constructed based on the analysis of efficient systolic arrays designs of the beamformer building blocks. The implemented overall structure is then validated to ensure its proper operation. Further, the implementation performance is evaluated in terms of accuracy and error analysis in comparison to the MATLAB simulations. The new methodology is based on the systematic methodology to close the gap between the modern wideband radar I/O rates and the silicon operating speed. This new metodology is applied to the interpolator block as an example. The proposed methodology is simulated and tested using MATLAB object oriented programming (OOP) to ensure the proper operation.
Analysis of algorithms for filter bank design optimization
(2019-09-06) ElGarewi, Ahmed; Agathoklis, Panajotis
This thesis deals with design algorithms for filter banks based on optimization. The design specifications consist of the perfect reconstruction and frequency response specifications for finite impulse response (FIR) analysis and synthesis filters. The perfect reconstruction conditions are formulated as a set of linear equations with respect to the analysis filters’ coefficients and the synthesis filters’ coefficients. Five design algorithms are presented. The first three are based on an unconstrained optimization of performance indices, which include the perfect reconstruction error and the error in the frequency specifications. The last two algorithms are formulated as constrained optimization problems with the perfect reconstruction error as the performance index and the frequency specifications as constraints. The performance of the five algorithms is evaluated and compared using six examples; these examples include uniform filter bank, compatible non-uniform filter bank and incompatible non-uniform filter bank designs. The evaluation criteria are based on distortion and aliasing errors, the magnitude response characteristics of analysis and synthesis filters, the computation time required for the optimization, and the convergence of the performance index with respect to the number of iterations. The results show that the five algorithms can achieve almost perfect reconstruction and can meet the frequency response specifications at an acceptable level. In the case of incompatible non-uniform filter banks, the algorithms have challenges to achieve almost perfect reconstruction.
Array Signal Processing for Beamforming and Blind Source Separation
(2013-04-30) Moazzen, Iman; Agathoklis, Panajotis
A new broadband beamformer composed of nested arrays (NAs), multi-dimensional (MD) filters, and multirate techniques is proposed for both linear and planar arrays. It is shown that this combination results in frequency-invariant response. For a given number of sensors, the advantage of using NAs is that the effective aperture for low temporal frequencies is larger than in the case of using uniform arrays. This leads to high spatial selectivity for low frequencies. For a given aperture size, the proposed beamformer can be implemented with significantly fewer sensors and less computation than uniform arrays with a slight deterioration in performance. Taking advantage of the Noble identity and polyphase structures, the proposed method can be efficiently implemented. Simulation results demonstrate the good performance of the proposed beamformer in terms of frequency-invariant response and computational requirements. The broadband beamformer requires a filter bank with a non-compatible set of sampling rates which is challenging to be designed. To address this issue, a filter bank design approach is presented. The approach is based on formulating the design problem as an optimization problem with a performance index which consists of a term depending on perfect reconstruction (PR) and a term depending on the magnitude specifications of the analysis filters. The design objectives are to achieve almost perfect reconstruction (PR) and have the analysis filters satisfying some prescribed frequency specifications. Several design examples are considered to show the satisfactory performance of the proposed method. A new blind multi-stage space-time equalizer (STE) is proposed which can separate narrowband sources from a mixed signal. Neither the direction of arrival (DOA) nor a training sequence is assumed to be available for the receiver. The beamformer and equalizer are jointly updated to combat both co-channel interference (CCI) and inter-symbol interference (ISI) effectively. Using subarray beamformers, the DOA, possibly time-varying, of the captured signal is estimated and tracked. The estimated DOA is used by the beamformer to provide strong CCI cancellation. In order to alleviate inter-stage error propagation significantly, a mean-square-error sorting algorithm is used which assigns detected sources to different stages according to the reconstruction error at different stages. Further, to speed up the convergence, a simple-yet-efficient DOA estimation algorithm is proposed which can provide good initial DOAs for the multi-stage STE. Simulation results illustrate the good performance of the proposed STE and show that it can effectively deal with changing DOAs and time variant channels.
Complexity reduction using frequency masking technique for FIR filter design
(2021-12-01) Kumar, Sanjeev; Agathoklis, Panajotis
A digital filter is an essential part of many present day electronics devices and a common type of digital filter used is Finite Impulse Response (FIR) filters. FIR filters have several advantages over Infinite Impulse Response (IIR) filters, such as easy design of linear phase FIR filters, inherently stability and low round off noise sensitivity. On the other hand, FIR filters have the disadvantage that when the design specifications require a narrow transition band, the order of FIR designs increases rapidly. One of the common techniques to design narrow transition band FIR filter is the Frequency Masking (FRM) technique which is based on up-sampling and the use of masking filters. This leads to filters with lower order and decreased arithmetic complexity of the filter implementation. In this project the design of linear phase FIR filters using the FRM technique is compared with some of the common ways to design linear phase FIR filters using windows. Design of linear phase low pass FIR filters using conventional window method technique is highlighted with different design examples implemented using MATLAB. The window technique has limitation as it lacks flexibility because, in design both the peak passband (δp) and stopband (δs) ripples are considered approximately equals, so that the designer can’t make a passband ripple very small or a stopband attenuation very large. Further, the order of FIR filters designed using the window technique tends to increase fast when a narrow transition band is required. These design limitations can be overcome by using the frequency masking techniques which provides high selectivity with reduced arithmetic complexity. Linear phase low pass narrow bandwidth FRM FIR filter design are implemented in MATLAB and compared with the window method design in terms of filter order. Two masking filter techniques are being considered. In the first one, a masking filter technique is being used in cascade with the up-sampled model filter to design the linear phase low pass narrow transition bandwidth FIR filter. In the second, two masking filters are being used in a two channel configuration to achieve more general passbands. The comparison of these FIR filter designs indicates that the FRM technique leads to lower order FIR filters compared with the FIR filter designs using window methods at the cost of a slightly more complex design process. This reduction of filter order implies that the FRM techniques can be used to reduce arithmetic complexity for FIR filter implementations in applications requiring linear phase narrow transition bandwidth FIR filters.
Design and implementation of discrete type 1 and type 2 controllers with residual feedback for modal adaptive optics
(2009-11-16T17:35:36Z) Hampton, Peter John; Bradley, Colin; Agathoklis, Panajotis
This thesis deals with the design. implementation and experimental validation of digital controllers for an Adaptive Optics (AO) test bench. The AO model has been modeled using a discrete model and this model was validated using experimental measurements. The controllers developed consist of 1-D parallel controllers, one for each mode of the optical system. Two controllers were designed and analyzed for the system: (i) a Type I controller that exhibits improved bandwidth over standard AO controllers and (ii) a Type 2 controller that is capable of low frequency prediction and can provide exceptionally low following error if the quantization noise sources are sufficiently small. The implementation of the controllers on the AO test bench was considered and the problems arising from the quantization of the output of the controller due to hardware limitations were studied. Controller implementations that improve the performance of the system with respect to stability despite the non-linear effects of the quantizer are provided. The controllers developed were implemented and their performance was evaluated and compared with the results expected from theoretical analysis. The experimental results are very close to the ones expected from the theoretical analysis except when the quantization effects become significant.
Design, Simulation and Implementation of High Precision Control Algorithms for a Galvanometer Laser Scanner
(2014-08-26) Torres Bonet, Tomas; Agathoklis, Panajotis
This thesis focuses on the theory, design, simulation and implementation of several digital controllers for periodic signals on a laser scanning galvanometer. A model for the galvanometer was obtained and veri ed using closed loop identi cation techniques. Using this model, controllers were designed and simulated using MATLAB and then implemented on a custom FPGA control processor with a focus on tracking performance. The types of controllers used were: an Iterative Learning Controller, an RST pole placement controller, an Adaptive Feed-forward cancellation controller, a combined Iterative Learning and Adaptive Feed-forward cancellation controller and a simple PID controller. The simulated results were better than the experimental results because of system noise and modelling uncertainties but the relative performance between each of the controllers was similar for both the simulation and experimental setup. The experimental results achieved were very good with one controller reaching errors under 50 rad.
Direction of arrival estimation technique for narrow-band signals based on spatial Discrete Fourier Transform
(2018-08-24) Zaeim, Ramin; Agathoklis, Panajotis
This work deals with the further development of a method for Direction of Arrival (DOA) estimation based on the Discrete Fourier Transform (DFT) of the sensor array output. In the existing DFT-based algorithm, relatively high SNR is considered, and it is assumed that a large number of sensors are available. In this study an overview of some of the most commonly used DOA estimation techniques will be presented. Then the performance of the DFT method will be analyzed and compared with the performance of existing techniques. Two main objectives will be studied, firstly the reduction of the number of sensors and secondly the performance of the DFT based technique in the presence of noise. Experimental simulations will be presented to illustrate that in absence of noise, the proposed method is very fast and using just one snapshot is sufficient to accurately estimate DOAs. Also, in presence of noise, the method is still relatively fast and using a small number of snapshots, it can accurately estimate DOAs. The above mentioned properties are the result of taking an average of the peaks of the DFTs, X_n (k), obtained from a sequence of N_s snapshots. With N_s sufficiently large, the average over N_s snapshots approaches expected value. Also, the conditions that should be satisfied to avoid overlapping of main-lobes, and thus loosing the DOA of some signals, in the DFT spectrum are examined. This study further analyzes the performance of the proposed method as well as two other commonly used algorithms, MUSIC and conventional beamformer. An extensive simulation was conducted and different features of the spatial DFT technique, such as accuracy, resolution, sensitivity to noise, effect of multiple snapshots and the number of sensors were evaluated and compared with those of existing techniques. The simulations indicate that in most aspects the proposed spatial DFT algorithm outperforms the other techniques.
A Distributed Intelligent Lighting Solution and the Design and Implementation of a Sensor Middleware System
(2015-04-30) Fischer, Michael; Wu, Kui; Agathoklis, Panajotis
This thesis addresses a multi-phase research and development project that spanned nearly four years, targeted at providing an ultra high-efficiency, user-friendly, and economic intelligent lighting solution for commercial facility applications, initially targeting underground parking specifically. The system would leverage the strengths of four key technologies: high brightness white Light Emitting Diodes (LEDs), wireless sensor and actuator networks, single board computers, and cloud computing. An introduction to these technologies and an overview of how they were combined to build an intelligent lighting solution is given, followed by an in-depth description of the design and implementation of one of the main subsystems – the Sensor Middleware System – residing on a single board computer. Newly-available LED luminaires (a.k.a. light fixtures) bring the combination of high efficiency, reliability, illumination quality, and long-lifetime to the lighting market. Emerging low-power – and recently low-cost – 802.15.4 wireless networks offer high controllability and responsiveness to deployed luminaires and sensors. The cost- associativity, low maintenance, and easy build-up of Internet Data Center “cloud” computing resources make data collection and remote management infrastructure for Building Automation Systems accessible to even small companies. Additionally, these resources can be much more appropriately sized and allocated, which reduces energy use. These technologies are combined to form an Intelligent Lighting System (ILS). Fitting well within the Internet of Things paradigm, this highly distributed messaging-based “system of systems” was designed to be reliable through loose coupling – spanning multiple network layers and messaging protocols. Its goal was to deliver significant energy savings over incumbent technologies, configurable and responsive lighting service behaviour, and improved experience for users within the facility (pedestrians and drivers) and those interacting with its web-based tools (building managers and ILS administrators). The ILS was partitioned into three main subsystems as follows. The installed Wireless Field Network (WFN) of luminaires and sensors provided coordinated scheduled and real-time output level adjustment (i.e. dimming), with the help of motion sensor triggers. The Monitoring and Configuration System (MCS) in the cloud provided remote data collection and a web-based monitoring and configuration Graphical User Interface application. Network hardware and Message-Oriented Middleware (MOM) were responsible for tying these subsystems together. The MOM layer that provided the message brokering, translating, envelope wrapping, and guaranteed delivery services between the WFN and MCS, as well as field supervisory and quality-of-service functions for the WFN, was called the Sensor Middleware System (SMS). It was hosted on a single board computer located at the facility.
Fusion of multi-exposure videos in the gradient domain
(2025) Rawat, Aryan; Agathoklis, Panajotis; Sima, Mihai
Gradient-domain processing has proven effective for image fusion tasks by preserving local contrast and structural details while avoiding intensity-domain artifacts. However, most existing fusion methods are designed for still images and fail to address temporal consistency when applied frame-by-frame to video sequences, often resulting in flicker and temporal instability. This seminar presents a novel approach for multi-exposure video fusion formulated directly in the three-dimensional gradient domain. The proposed method operates on spatial-temporal gradients extracted from registered exposure sequences and fuses them using a gradient-selection strategy. Reconstruction of the fused video is performed using a wavelet-based 3-D Haar decomposition combined with an iterative Poisson solver, ensuring both spatial fidelity and temporal coherence. Experimental results on standard video datasets demonstrate that the proposed method effectively enhances visual detail while significantly reducing temporal artifacts compared to classical intensity-based fusion techniques. Quantitative evaluation using spatial and temporal metrics further confirms the advantages of gradient-domain fusion for video applications.
Image stitching and object insertion in the gradient domain
(2011-12-20) Sevcenco, Ioana Speranta; Agathoklis, Panajotis
In this thesis, the applications of image stitching and object insertion are considered and two gradient based approaches offering solutions are proposed. An essential part of the proposed methods is obtaining an image from a given gradient data set. This is done using an existing Haar wavelet based reconstruction technique, which consists of two main steps. First, the Haar wavelet decomposition of the image to be reconstructed is obtained directly from a given gradient. Second, the image is obtained using Haar wavelet synthesis. In both stitching and object insertion applications considered, the gradient from which the image must be reconstructed is a non-conservative vector field and this requires adding an iterative Poisson solver at each resolution level, during the synthesis step of the reconstruction technique. The performance of the reconstruction algorithm is evaluated by comparing it with other existing techniques, in terms of solution accuracy and computation speed. The proposed image stitching technique consists of three main parts: registering the images to be combined, blending their gradients over a region of interest and obtaining a composite image from a gradient. The object insertion technique considers the images registered and has two main stages: gradient blending of images in a region of interest and recovering an image from the gradient. The performance of the stitching algorithm is evaluated visually, by presenting the results produced to combine images with varying orientation, scales, illumination, and color conditions. Experimental results illustrate both the stitching and the insertion techniques proposed, and indicate that they yield seamless composite images.
Implementation and performance analysis of 3D cone and frustum filters
(2015-08-07) Shubayli, Hussam; Reddy, Hari; Agathoklis, Panajotis
In this thesis, new effective and efficient implementation structures of three-dimensional (3D) spatio-temporal (ST) Finite Impulse Response (FIR) uniform and non-uniform cone and frustum filters using well-known filter banks are investigated. The performance of the proposed implementation structures for 3D ST FIR uniform and non-uniform cone and frustum filters are investigated for 3D broadband beamforming in radio astronomy applications. First, implementations of two 3D ST uniform FIR cone filters are investigated. The 3D cone filters are designed by cascading either the well-known uniform quadrature mirror – cosine-modulated (QM-CM) filter bank or directly designed filter banks (DDFBs), with 2D low-pass circularly-symmetric spatial filters. In addition, two 3D ST uniform FIR frustum filters are derived from the cone filters by implementing partial bands of the filter banks with corresponding 2D spatial filters. The performance of the proposed implementation structures for 3D ST uniform QM-CM and DDFBs cone and frustum filters are evaluated using broadband beamforming signals in radio astronomy applications. The performance of the QM-CM and DDFBs cone and frustum filters shows improvement in terms of Signal-to-Interference-plus-Noise ratio (SINRs) over existing 3D ST cone and frustum filters. In addition to their effective performance, these cone and frustum filters can be efficiently implemented with equivalent or less computational complexity compared to existing methods. Second, implementations of two 3D ST non-uniform cone and frustum filters are explored. These cone and frustum filters are obtained by cascading either QM-CM or DDFBs non-uniform filter banks, with 2D low-pass circularly-symmetric spatial filters. The motivation for the 3D ST non-uniform cone and frustum filters is to achieve better approximation at low temporal frequencies than using the uniform ones. The performance of the 3D ST non-uniform cone and frustum filters is evaluated and compared with the performance of the uniform 3D ST cone and frustum filters. Results indicate that the performance of the proposed 3D ST non-uniform QM-CM and DDFBs cone filters shows some improvement in selective filtering compared to the performance of 3D ST uniform cone filters.
Infrared-Visible Image Fusion in the Gradient Domain
(2024) Premaratne, Sanduni; Agathoklis, Panajotis; Bruton, Leonard T.
Due to the complementary properties of the infrared cameras compared to conventional visible imaging cameras, it has become increasingly popular to fuse infrared and visible images of the same scene for better visual understanding. One major application of this is surveillance which involves videos and requires fast processing. Therefore, there is a need for investigating novel low-complexity fusion algorithms that can be implemented in real-time applications. In this study, we address this critical research problem by two-scale fusion in the gradient domain with saliency detection and image enhancement. In the proposed method, the source images are first decomposed in to base and detail layers. Next, the base parts are fused in the gradient domain by choosing the maximum absolute gradient, whereas the gradients of the detail parts are fused using a weighted average where the weights are calculated using saliency maps. Prior to fusion, the detail parts are enhanced using a guided filter-based enhancement approach. Finally, the fused gradients of the base and detail components are added together to obtain the gradients of the fused image, from which the fused image is reconstructed using a reconstruction technique based on wavelets. Experimental results demonstrate that the proposed method achieves very competitive performance in subjective and objective fusion assessments, while also outperforming most methods in terms of computational complexity.
Low-Complexity Multi-Dimensional Filters for Plenoptic Signal Processing
(2015-12-02) Edussooriya, Chamira Udaya Shantha; Bruton, Leonard T.; Agathoklis, Panajotis
Five-dimensional (5-D) light field video (LFV) (also known as plenoptic video) is a more powerful form of representing information of dynamic scenes compared to conventional three-dimensional (3-D) video. In this dissertation, the spectra of moving objects in LFVs are analyzed, and it is shown that such moving objects can be enhanced based on their depth and velocity by employing 5-D digital filters, what is defined as depth-velocity filters. In particular, the spectral region of support (ROS) of a Lambertian object moving with constant velocity and at constant depth is shown to be a skewed 3-D hyperfan in the 5-D frequency domain. Furthermore, it is shown that the spectral ROS of a Lambertian object moving at non-constant depth can be approximated as a sequence of ROSs, each of which is a skewed 3-D hyperfan, in the 5-D continuous frequency domain. Based on the spectral analysis, a novel 5-D finite-extent impulse response (FIR) depth-velocity filter and a novel ultra-low complexity 5-D infinite-extent impulse response (IIR) depth-velocity filter are proposed for enhancing objects moving with constant velocity and at constant depth in LFVs. Furthermore, a novel ultra-low complexity 5-D IIR adaptive depth-velocity filter is proposed for enhancing objects moving at non-constant depth in LFVs. Also, an ultra-low complexity 3-D linear-phase IIR velocity filter that can be incorporated to design 5-D IIR depth-velocity filters is proposed. To the best of the author’s knowledge, the proposed 5-D FIR and IIR depth-velocity filters and the proposed 5-D IIR adaptive depth-velocity filter are the first such 5-D filters applied for enhancing moving objects in LFVs based on their depth and velocity. Numerically generated LFVs and LFVs of real scenes, generated by means of a commercially available Lytro light field (LF) camera, are used to test the effectiveness of the proposed 5-D depth-velocity filters. Numerical simulation results indicate that the proposed 5-D depth-velocity filters outperform the 3-D velocity filters and the four-dimensional (4-D) depth filters in enhancing moving objects in LFVs. More importantly, the proposed 5-D depth-velocity filters are capable of exposing heavily occluded parts of a scene and of attenuating noise significantly. Considering the ultra-low complexity, the proposed 5-D IIR depth-velocity filter and the proposed 5-D IIR adaptive depth-velocity filter have significant potentials to be employed in real-time applications.
Multi-dimensional digital signal integration with applications in image, video and light field processing
(2018-08-16) Sevcenco, Ioana Speranta; Agathoklis, Panajotis
Multi-dimensional digital signals have become an intertwined part of day to day life, from digital images and videos used to capture and share life experiences, to more powerful scene representations such as light field images, which open the gate to previously challenging tasks, such as post capture refocusing or eliminating visible occlusions from a scene. This dissertation delves into the world of multi-dimensional signal processing and introduces a tool of particular use for gradient based solutions of well-known signal processing problems. Specifically, a technique to reconstruct a signal from a given gradient data set is developed in the case of two dimensional (2-D), three dimensional (3-D) and four dimensional (4-D) digital signals. The reconstruction technique is multiresolution in nature, and begins by using the given gradient to generate a multi-dimensional Haar wavelet decomposition of the signals of interest, and then reconstructs the signal by Haar wavelet synthesis, performed on successive resolution levels. The challenges in developing this technique are non-trivial and are brought about by the applications at hand. For example, in video content replacement, the gradient data from which a video sequence needs to be reconstructed is a combination of gradient values that belong to different video sequences. In most cases, such operations disrupt the conservative nature of the gradient data set. The effects of the non-conservative nature of the newly generated gradient data set are attenuated by using an iterative Poisson solver at each resolution level during the reconstruction. A second and more important challenge is brought about by the increase in signal dimensionality. In a previous approach, an intermediate extended signal with symmetric region of support is obtained, and the signal of interest is extracted from it. This approach is reasonable in 2-D, but becomes less appealing as the signal dimensionality increases. To avoid generating data that is then discarded, a new approach is proposed, in which signal extension is no longer performed. Instead, different procedures are suggested to generate a non-symmetric Haar wavelet decomposition of the signals of interest. In the case of 2-D and 3-D signals, ways to obtain this decomposition exactly from the given gradient data and the average value of the signal are proposed. In addition, ways to approximate a subset of decomposition coefficients are introduced and the visual consequences of such approximations are studied in the special case of 2-D digital images. Several ways to approximate the same subset of decomposition coefficients are developed in the special case of 4-D light field images. Experiments run on various 2-D, 3-D and 4-D test signals are included to provide an insight on the performance of the reconstruction technique. The value of the multi-dimensional reconstruction technique is then demonstrated by including it in a number of signal processing applications. First, an efficient algorithm is developed with the purpose of combining information from the gradient of a set of 2-D images with different regions in focus or different exposure times, with the purpose of generating an all-in-focus image or revealing details that were lost due to improper exposure setting. Moving on to 3-D signal processing applications, two video editing problems are studied and gradient based solutions are presented. In the first one, the objective is to seamlessly place content from one video sequence in another, while in the second one, to combine elements from two video sequences and generate a transparency effect. Lastly, a gradient based technique for editing 4-D scene representations (light fields) is presented, as well as a technique to combine information from two light fields with the purpose of generating a light field with more details of the imaged scene. All these applications show that the developed technique is a reliable tool for gradient domain based solutions of signal processing problems.
Multidimensional signal processing techniques for disturbance mitigation in synthetic aperture systems
(2012-08-21) Edussooriya, Chamira Udaya Shantha; Bruton, Leonard T.; Agathoklis, Panajotis
In this thesis, multidimensional signal processing techniques to mitigate disturbances in synthetic aperture systems such as radio telescopes are investigated. Here, two computationally efficient three-dimensional (3D) spatio-temporal (ST) finite impulse response (FIR) cone filter bank structures are proposed. Furthermore, a strategy is proposed to design 3D ST FIR frustum filter banks, having double-frustum-shaped passbands oriented along the temporal axis, derived from appropriate 3D ST FIR cone filter banks. Both types of cone and frustum filter banks are almost alias free and provide near-perfect reconstruction. In the proposed cone and frustum filter banks, both temporal and spatial filtering operations can be carried out at a significantly lower rate compared to previously reported 3D ST FIR cone filter banks implying lower power consumption. Furthermore, the proposed cone and frustum filter banks require a significantly lower computational complexity than previously reported 3D ST FIR cone and frustum filter banks. Importantly, this is achieved without deteriorating the improvement in signal-to-interference-plus-noise ratio. A theoretical analysis of brightness distribution (BD) errors caused by parameter perturbations and mismatches among the transfer functions of receivers employed in synthetic aperture systems is presented. First, the BD errors caused by perturbations in the transfer functions of low noise amplifiers (LNAs) and anti-aliasing filters (AAFs) are considered, and the characteristics of the additive BD error and its effects on synthesized BDs are thoroughly analyzed. Second, the conditions that should be satisfied by the transfer functions of digital beamformers to eliminate the BD errors caused by their phase responses are examined. The sufficient condition to eliminate the BD errors is that the transfer functions are matched, and, interestingly, the phase responses are not necessary to be linear. Furthermore, the BD errors caused by typical tolerances of passive L and C elements used to implement the AAFs and those caused by the random variations of gain from LNA to LNA are quantified through numerical simulations. The simulations indicate that substantial BD errors are observed at frequencies that are close to the passband edge of the AAFs.
New beamforming and DOA estimation techniques in wireless communications
(2008-04-10T06:04:08Z) Wang, Nanyan.; Agathoklis, Panajotis; Antoniou, Andreas
The development and performance evaluation of new techniques for direction of arrival (DOA) estimation, single-user beamforming (SUB), and multiuser beamforming (MUB) to be used in wireless communications are investigated. Two of the most commonly used beamformer configurations in direct-sequence code-division multiple access (DS-CDMA) systems, the chip-based (CB) and the symbol-based (SB) configurations for the base station (BS) receiver, are studied and their performance is evaluated. It is shown that using the CB configuration, different interfering components are rejected based on the spatial distribution of their power. In the SB configuration, spatial diversity is exploited after despreading and different interfering components are rejected based on their interfering strength which depends on both their power and code correlation with the signal of interest. For the SB configuration, more effort is applied to rejecting the interfering components with higher interfering strength and thus a more selective and efficient system is achieved. Detailed performance analysis and simulations show that in the presence of multipleaccess interference, the SB configuration can lead to a significant improvement in the signal-to-interference-plus-noise ratio relative to that achieved with the CB configuration for both asynchronous and synchronous DS-CDMA systems. A new technique for DOA estimation is proposed. The new technique, called subarray beamforming-based DOA (SBDOA) estimation, uses two virtual subarrays to form a signal whose phase relative to the reference signal is a function of the DOA. The DOA is then estimated based on the computation of the phase shift between the reference signal and the phaseshifted target signal. Since the phase-shifted target signal is obtained after interference rejection through beamforrning, the effect of co-channel interference on the estimation is significantly reduced. The proposed technique is computationally simple and the number of signal sources detectable is iii not bounded by the number of antenna elements used. Performance analysis and extensive simulations show that the proposed technique offers significantly improved estimation resolution, capacity, accuracy, and tracking capability relative to existing techniques. A new SUB algorithm is proposed for the downlink in wireless communication systems. The beam pattern at the BS is determined using a new optimality criterion which takes into consideration the fact that signals from the BS to different mobile stations (MSs) have different power and thus have different resistance to co-channel interference. In this way, the power of co-channel interference in the direction of an MS whose downlink signal has low resistance to co-channel interference can be significantly reduced. Simulation results show that the new algorithm leads to better performance than conventional algorithms in terms of system outage probability. A new MUB algorithm is proposed for joint beamforming and power control for the downlink in wireless communication systems. The optimization problem of optimal MUB is reformulated by modifying the constraints so that the weight vectors at the BS for different MSs are optimized in a feasible region which is a subset of the one of the original MUB problem. The downlink beamforming weight vectors of different MSs are then jointly optimized in a subspace instead of searching in the entire parameter space. Simulation results show that the modified optimization problem leads to solutions that satisfy the signal-to-noise-plus-interference ratio specification at each MS and, at the same time, the total power transmitted from the BS is very close to the optimal one. The solution of the modified optimization problem requires significantly less computation than that of the optimal MUB problem.
New efficient block-based motion estimation algorithms for video compression and their hardware implementations
(2010-02-04T19:13:45Z) Rehan, Mohamed Mohamed; Agathoklis, Panajotis; Antoniou, Andreas
Video compression technology aims at compressing large amount of video data for efficient transmission and storage without significant loss of quality. Most video compression techniques rely on removing temporal data redundancy between frames using motion estimation and motion compensation techniques which are generally very computationally expensive. The objective of the research done in this thesis is to develop new efficient motion estimation techniques that reduce the computational complexity of motion estimation. The thesis presents a new prediction technique referred to as weighted sum block matching (WSBM) which dynamically reduces the computational complexity by limiting the search to a small subset of the search area. Simulation results have shown that adding WSBM to some well-known search algorithms reduces their computational complexity by 6-1.5 without affecting the visual quality of the reconstructed video frames. The thesis also presents two new algorithms based on the simplex optimization method. the simplex based block matching algorithm (SMPLX) and the flexible triangle search (FTS). Both techniques use a triangle that moves inside the search area and checks only positions that lie at its vertices. As a result the computational complexity of the search is reduced since it depends directly on the number of positions checked. The techniques can change the size and orientation of the search triangle during the search. The changes make the search highly flexible and efficient and reduce the number of search positions to be checked compared to those in other search algorithms. The SMPLX uses equations based on the simplex optimization method to compute the new triangle size and orientation. The FTS, on the other hand, was implemented to be more suitable for a digital search grid by using look-up tables and integer computations. The two algorithms were implemented as part of the H.263 and H.264 encoders. Both algorithms were compared to the state of the art motion search algorithms. Experimental results showed that both algorithms can reach sub-optimal solutions while checking fewer search positions compared to other algorithms which results in lower computational complexity as a consequence. Additional research was done to analyze and further improve FTS performance. As a result, various extensions of the FTS have been developed such as the enhanced FTS (EFTS), the half-pixel FTS (HP-FTS). and the predictive FTS (PETS). These extensions were also implemented as part of the H.263 and H.264 encoders. In the EFTS. repeated computations are reduced by caching intermediate results. In addition. the termination condition is modified to avoid premature exit. These modifications reduce the computational complexity of the FTS by up to 4%%. The HP-FTS extended the FTS so that the search can be done at half-pixel resolution instead of full-pixel resolution. The commonly used approach for half-pixel search is based on two separate stages. i.e., full-pixel search followed by half-pixel search. By combining the two stages in HP-FTS. the overall computational complexity can be reduced by an average of 13% without affecting the produced quality or compression ratio. The PETS uses prediction to select the direction of the starting search triangle. Analysis results show that the proper selection of the starting search triangle has great effect on the performance of the FTS. Simulation results show that the PFTS can reduce the computational complexity of the FTS by 7-13%. Finally, hardware designs for the FTS and the full search (FS) algorithms are proposed. The FS was chosen due to its regularity, low control overhead, and suitability for hardware implementation. It uses a high degree of parallelism and pipelining in order to improve the computational efficiency. The FTS requires less computation and thus provides high processing rates. Both designs were implemented, simulated, and verified using VHDL and then synthesized with Xilinx FPGAs. Simulation results have shown that both hard-ware implementations are more efficient than other existing implementations in terms of performance and hardware usage.

Browsing by Supervisor "Agathoklis, Panajotis"

Results Per Page

Sort Options