Robert E. Bogner

An insect vision-based motion detection chip

The architectural and circuit design aspects of a mixed analog/digital very large scale integration (VLSI) motion detection chip based on models of the insect visual system are described. The chip comprises two one-dimensional 64-cell arrays as well as front-end analog circuitry for early visual processing and digital control circuits. Each analog processing cell comprises a photodetector, circuits for spatial averaging and multiplicative noise cancellation, differentiation, and thresholding. The operation and configuration of the analog cells is controlled by digital circuits, thus implementing a reconfigurable architecture which facilitates the evaluation of several newly designed analog circuits. The chip has been designed and fabricated in a 1.2-/spl mu/m CMOS process and occupies an area of 2/spl times/2 mm/sup 2/.

Dual /spl nu/-support vector machine with error rate and training size biasing

Support vector machines (SVMs) have been successfully applied to classification problems. The difficulty in selecting the most effective error penalty has been partly resolved with /spl nu/-SVM. However, the use of uneven training class sizes, which occurs frequently with target detection problems, results in machines with biases towards the class with the larger training set. We propose an extended /spl nu/-SVM to counter the effects of the unbalanced training class sizes. The resulting dual /spl nu/-SVM provides the facility to counter these effects, as well as to adjust the error penalties of each class separately. The parameter /spl nu/ of each class provides a lower bound to the fraction of support vector of that class, and the upper bound to the fraction of bounded support vector of that class. These bounds allow the control on the error rates allowed for each class, and enable the training of machines with specific error rate requirements.

Blind Signal Separation I. Linear, Instantaneous Combinations: I. Linear, Instantaneous Combinations

Abstract Pope, K. J., and Bogner, R. E., Blind Signal Separation. I. Linear, Instantaneous Combinations,Digital Signal Processing6, 5–16. Blind signal separation is the process of extracting unknown independent source signals from sensor measurements which are unknown combinations of the source signals. The term “blind” is used as the source signals and the method of combination are unknown, and hence the problem is related to the problems of blind deconvolution and blind equalization. Blind signal separation is sometimes referred to as independent component analysis (InCA), as it generalizes principal component analysis to produce independent signals rather than simply uncorrelated signals. The problem of blind signal separation has been investigated in detail during the past ten years. The work has been driven by a wide variety of interests and areas of application, such as array beam-forming, higher-order statistics, neural networks and artificial learning, noise cancellation, and speech enhancement. However, no review of the available literature has been published. This paper is one of two papers seeking to redress this point. It focuses on the simplest form of the blind separation problem: the separation of sources that have combined in a linear, instantaneous fashion.

Blind Signal Separation II. Linear, Convolutive Combinations: II. Linear, Convolutive Combinations

Abstract Pope, K. J., and Bogner, R. E., Blind Signal Separation. II. Linear, Convolutive Combinations, Digital Signal Processing 6 , 17–28. Blind signal separation is the process of extracting unknown independent source signals from sensor measurements which are unknown combinations of the source signals. The term “blind” is used as the source signals and the method of combination are unknown, and hence the problem is related to the problems of blind deconvolution and blind equalization. Blind signal separation is sometimes referred to as independent component analysis (InCA), as it generalizes principal component analysis to produce independent signals rather than simply uncorrelated signals. The problem of blind signal separation has been investigated in detail during the past ten years. The work has been driven by a wide variety of interests and areas of application, such as array beam-forming, higher-order statistics, neural networks and artificial learning, noise cancellation, and speech enhancement. However, no review of the available literature has been published. This paper is one of two papers seeking to redress this point. Whereas Part I focused on the separation of sources that have combined in a linear, instantaneous fashion, this paper considers a more complicated separation problem in which the combination of the sources is linear and convolutive. In addition, a variety of issues of importance to blind signal separation problems in general are also discussed.

Three-dimensional space-borne synthetic aperture radar (SAR) imaging with multiple pass processing

Three-dimensional (3D) synthetic aperture radar (SAR) imaging via multiple-pass processing is an extension of interferometric SAR imaging. It exploits more than two flight passes to achieve a desired resolution in elevation. In this paper, a novel approach is developed to reconstruct a 3D space-borne SAR image with multiple-pass processing. It involves image registration, phase correction and elevational imaging. An image model matching is developed for multiple image registration, an eigenvector method is proposed for the phase correction and the elevational imaging is conducted using a Fourier transform or a super-resolution method for enhancement of elevational resolution. 3D SAR images are obtained by processing simulated data and real data from the first European Remote Sensing satellite (ERS-1) with the proposed approaches.

An Implementation of Training Dual-nu Support Vector Machines

Dual-ν Support Vector Machine (2ν-SVM) is a SVM extension that reduces the complexity of selecting the right value of the error parameter selection. However, the techniques used for solving the training problem of the original SVM cannot be directly applied to 2ν-SVM. An iterative decomposition method for training this class of SVM is described in this chapter. The training is divided into the initialisation process and the optimisation process, with both processes using similar iterative techniques. Implementation issues, such as caching, which reduces the memory usage and redundant kernel calculations are discussed.

An analog implementation of early visual processing in insects

An analog VLSI implementation which mimics the early visual processing stages in insects is described. The system is composed of sixty parallel channels of integrated photodetectors and processing elements. It serves as the front end processor for a motion detection chip. The photodetection circuitry includes p-well junction diodes on a 2 mu m CMOS process and a logarithmic compression to increase the dynamic range of the system. The processing elements consist of an analog differentiator behind each photodetector. The differentiators are low frequency and have been designed using subthreshold design methods.<<ETX>>

Three-dimensional SAR imaging via multiple pass processing

This paper develops a novel approach to reconstruct a three-dimensional (3D) SAR image with multiple pass processing. It involves image registration, phase correction and beamforming in elevation. An eigenvector method is proposed for the phase correction and the beamforming in elevation is carried out by a DFT or a subspace method for superresolution. 3D SAR images are demonstrated by processing ERS-1 real data with the proposed approach.

Radar target recognition using range profiles

With the increased availability of coherent wideband radars, there has been a renewed interest in radar target recognition. A large bandwidth gives high resolution in range which means target recognition may be possible. We examine some of the problems of classifying high resolution range profiles (HRRP), and investigate simple preprocessing techniques which may aid subsequent target classification. We apply these techniques to HRRP data acquired at a local airport using the Microwave Radar Division (MRD) mobile radar facility. We find that we can reliably distinguish between Boeing 727 and Boeing 737 aircraft over a range of aspect angles.<<ETX>>

High-resolution radar imagery of the Mirage III aircraft

High-resolution radar imagery has attracted increasing interest in recent years. As more radars are endowed with a high-resolution capability, target classification will become a regular system function. In order to classify an aircraft using radar, one must have an understanding of how the radar imagery relates to the physical aircraft. This paper illustrates the more important radar backscattering features on a typical fighter aircraft. Radar backscatter from an aircraft can occur through a variety of mechanisms. Although direct specular and diffractive mechanisms usually account for the majority of the scattering, indirect phenomena such as cavity scattering and creeping wave scattering are significant. This investigation finds that scattering from engine cavities is a particularly important radar backscatter mechanism for a fighter aircraft. Radar data of an actual Mirage aircraft is collected from a target turntable facility. This data is processed to obtain high-resolution range profiles (HRRP) and inverse synthetic aperture radar (ISAR) images, which indicate the prominent radar scatterers on the aircraft. The imagery is qualitatively examined, and its suitability for target classification is discussed. >