A numerical study of two classes of neural network models is presented. The performance of Ising spin neural networks as content-addressable memories for the storage of bit patterns is analysed. By studying systems of increasing sizes, behaviour consistent with fintite-size scaling, characteristic of a first-order phase transition, is shown to be exhibited by the basins of attraction of the stored patterns in the Hopfield model. A local iterative learning algorithm is then developed for these models which is shown to achieve perfect storage of nominated patterns with near-optimal content-addressability. Similar scaling behaviour of the associated basins of attraction is observed. For both this learning algorithm and the Hopfield model, by extrapolating to the thermodynamic limit, estimates are obtained for the critical minimum overlap which an input pattern must have with a stored pattern in order to successfully retrieve it. The role of a neural network as a tool for optimising cost functions of binary valued variables is also studied. The particular application considered is that of restoring binary images which have become corrupted by noise. Image restorations are achieved by representing the array of pixel intensities as a network of analogue neurons. The performance of the network is shown to compare favourably with two other deterministic methods-a gradient descent on the same cost function and a majority-rule scheme-both in terms of restoring images and in terms of minimising the cost function. All of the computationally intensive simulations exploit the inherent parallelism in the models: both SIMD (the ICL DAP) and MIMD (the Meiko Computing Surface) machines are used.

No description available.

Certain nonlinear binary codes having at least twice as many codewords as any known linear binary code can be regarded as the binary images of linear codes over Z4. This vision leads to a new concept in coding theory, called the Z4-linearity of binary codes. This thesis is a survey on the linear quaternary codes and their binary images under the Gray map. The conditions for the binary image of a linear quaternary code to be linear are thoroughly investigated and the Z4-linearity of the Reed-Muller and Hamming codes is discussed. The contribution of this study is a simplification on the testing method of linearity conditions via a few new lemmas and propositions. Moreover, binary images (of length 8) of all linear quaternary codes of length 4 are analyzed and it is shown that all 184 binary codes in the nonlinear subset of these images are worse than the (8, 4) Hamming code. This thesis also includes the Hensel lift and Galois ring which are important tools for the study of quaternary cyclic codes. Accordingly, the quaternary cyclic versions of the well-known nonlinear binary codes such as the Kerdock and Preparata codes and their Z4-linearity are studied in detail.

Multiframe image superresolution has been an active research area for many years. In this approach image processing techniques are used to combine multiple low-resolution (LR) images capturing different views of an object. These multiple images are generally under-sampled, degraded by optical and pixel blurs, and corrupted by measurement noise. We exploit diversities in the imaging channels, namely, the number of cameras, magnification, position, and rotation, to undo degradations. Using an iterative back-projection (IBP) algorithm we quantify the improvements in image fidelity gained by using multiple frames compared to single frame, and discuss effects of system parameters on the reconstruction fidelity. As an example, for a system in which the pixel size is matched to optical blur size at a moderate detector noise, we can reduce the reconstruction root-mean-square-error by 570% by using 16 cameras and a large amount of diversity in deployment.We develop a new technique for superresolving binary imagery by incorporating finite-alphabet prior knowledge. We employ a message-passing based algorithm called two-dimensional distributed data detection (2D4) to estimate the object pixel likelihoods. We present a novel complexity-reduction technique that makes the algorithm suitable even for channels with support size as large as 5x5 object pixels. We compare the performance and complexity of 2D4 with that of IBP. In an imaging system with an optical blur spot matched to pixel size, and four 2x2 undersampled LR images, the reconstruction error for 2D4 is 300 times smaller than that for IBP at a signal-to-noise ratio of 38dB.We also present a transform-domain superresolution algorithm to efficiently incorporate sparsity as a form of prior knowledge. The prior knowledge that the object is sparse in some domain is incorporated in two ways: first we use the popular L1 norm as the regularization operator. Secondly we model wavelet coefficients of natural objects using generalized Gaussian densities. The model parameters are learned from a set of training objects and the regularization operator is derived from these parameters. We compare the results from our algorithms with an expectation-maximization (EM) algorithm for L1 norm minimization and also with the linear minimum mean squared error (LMMSE) estimator.

Autonomous vehicles are the subject of intense research because they are a safe and convenient alternative to present-day vehicles. Human drivers base their navigational decisions primarily on visual information and researchers have been attempting to use computers to do the same. The current challenge in using computer vision lies not in the collection or transmission of visual data, but in the perception of visual data to extract from it useful information. The focus of this thesis is on the use of computer vision to navigate an autonomous vehicle that will participate in the Intelligent Ground Vehicle Competition (IGVC.) This document starts with a description of the IGVC and the software design of an autonomous vehicle. This thesis then focuses on the weakest link in the system - the computer vision module. Vehicles at the IGVC are expected to autonomously navigate an obstacle course. Competing vehicles need to recognize and stay between lines painted on grass or pavement. The research presented in this document describes two methods used for boundary line extraction: color-based object extraction, and shape analysis for line recognition. This is the first time a combination of these methods is being applied to the problem of line recognition in the context of the IGVC. The most significant contribution of this work is a method for extracting lines in a binary image even when the line is attached to a shape that is not a line. Novel methods have been used to simplify camera calibration, and for perspective correction of the image. The results give promise of vastly improved autonomous vehicle performance. / Master of Science

M.Phil. / This work proposes a fast local image feature detector and descriptor that is im- plementable on a GPU. The BFROST feature detector is the first published GPU implementation of the popular FAST detector. A simple but novel method of feature orientation estimation which can be calculated in constant time is proposed. The robustness and reliability of the orientation estimation is validated against rotation invariant descriptors such as SIFT and SURF. Furthermore, the BFROST feature descriptor is robust to noise, scalable, rotation invariant, fast to compute in parallel and maintains low memory usage. It is demonstrated that BFROST is usable in real-time applications such as vision-based localization and mapping of images captured from micro aerial platforms.

The objective of the research presented in this thesis is to investigate alternative ways of information processing employing asynchronous, data driven, and analogue computation in massively parallel cellular processor arrays, with applications in machine vision and artificial intelligence. The use of cellular processor architectures, with only local neighbourhood connectivity, is considered in VLSI realisations of the trigger-wave propagation in binary image processing, and in Bayesian inference. Design issues, critical in terms of the computational precision and system performance, are extensively analysed, accounting for the non-ideal operation of MOS devices caused by the second order effects, noise and parameter mismatch. In particular, CMOS hardware solutions for two specific tasks: binary image skeletonization and sum-product algorithm for belief propagation in factor graphs, are considered, targeting efficient design in terms of the processing speed, power, area, and computational precision. The major contributions of this research are in the area of continuous-time and discrete-time CMOS circuit design, with applications in moderate precision analogue and asynchronous computation, accounting for parameter variability. Various analogue and digital circuit realisations, operating in the continuous-time and discrete-time domains, are analysed in theory and verified using combined Matlab-Hspice simulations, providing a versatile framework suitable for custom specific analyses, verification and optimisation of the designed systems. Novel solutions, exhibiting reduced impact of parameter variability on the circuit operation, are presented and applied in the designs of the arithmetic circuits for matrix-vector operations and in the data driven asynchronous processor arrays for binary image processing. Several mismatch optimisation techniques are demonstrated, based on the use of switched-current approach in the design of current-mode Gilbert multiplier circuit, novel biasing scheme in the design of tunable delay gates, and averaging technique applied to the analogue continuous-time circuits realisations of Bayesian networks. The most promising circuit solutions were implemented on the PPATC test chip, fabricated in a standard 90 nm CMOS process, and verified in experiments.

Mathematical morphology stems from set theory and it makes use of properties of point sets. The first point set is an origin image and the second one (usually smaller) is a structuring element. Morphological image transformations are image to image transformations based on a few elementary set operators. Fundamental morphologic operations are dilation, erosion and hit or miss. Next operations described in this work are opening and closing. Originally morphological operators were used for binary images only, later they were generalized for grey tone and color ones. This work describes the basic morphological image processing methods including their practical usage in image filtering and segmentation.