Learning small gallery size for prediction of recognition performance on large populations

This paper addresses the estimation of a small gallery size that can generate the optimal error estimate and its confidence on a large population (relative to the size of the gallery) which is one of the fundamental problems encountered in performance prediction for object recognition. It uses a generalized two-dimensional prediction model that combines a hypergeometric probability distribution model with a binomial model and also considers the data distortion problem in large populations. Learning is incorporated in the prediction process in order to find the optimal small gallery size and to improve the prediction. The Chernoff and Chebychev inequalities are used as a guide to obtain the small gallery size. During the prediction, the expectation–maximization (EM) algorithm is used to learn the match score and the non-match score distributions that are represented as a mixture of Gaussians. The optimal size of the small gallery is learned by comparing it with the sizes obtained by the statistical approaches and at the same time the upper and lower bounds for the prediction on large populations are obtained. Results for the prediction are presented for the NIST-4 fingerprint database.     

Predictive models for multibiometric systems

Recognizing a subject given a set of biometrics is a fundamental pattern recognition problem. This paper builds novel statistical models for multibiometric systems using geometric and multinomial distributions. These models are generic as they are only based on the similarity scores produced by a recognition system. They predict the bounds on the range of indices within which a test subject is likely to be present in a sorted set of similarity scores. These bounds are then used in the multibiometric recognition system to predict a smaller subset of subjects from the database as probable candidates for a given test subject. Experimental results show that the proposed models enhance the recognition rate beyond the underlying matching algorithms for multiple face views, fingerprints, palm prints, irises and their combinations.     

BEES: exploring Mars with bioinspired technologies

The bioinspired engineering of exploration systems (BEES) applies insects' biological sensory and flight control abilities to the design of real-time autonomous, visual-navigation and control systems for small unmanned flying vehicles. The bioinspired engineering of exploration systems focuses on using the variety of nature-tested mechanisms successfully implemented by biological organisms but not easily accomplished by conventional methods. We apply BEES technology to the development of bioinspired visual navigation sensors integrated on small flyers to enable autonomous flight. BEES technology extracts the salient principles from a variety of diverse organisms adept at flight, and applies them to machines that can fly on Mars.     

Optically addressed ferroelectric memory with nondestructive readout

We present a review of the emerging optically addressed ferroelectric memory with nondestructive readout as a nonvolatile memory technology, identify its high-impact applications, and project on some novel device designs and architectures that will enable its realization. Based on the high-speed bidirectional polarization-dependent photoresponse, simulation of a readout circuit for a 16-kbit VLSI ferromemory chip yields read-access times of ~20 ns and read-cycle times of ~30 ns (~34 ns and ~44 ns, respectively, within a framework of a radiation-hard environment), easily surpassing those of the conventional electrical destructive readout. Extension of the simulation for a 64-kbit memory shows that the read-access and -cycle times are only marginally increased to ~21 ns and ~31 ns, respectively (~38 ns and ~48 ns, with a radiation-hard readout circuitry). Commercial realization of the optical nondestructive readout, however, would require a reduction in the incident (optical) power by roughly an order of magnitude for the readout or an enhancement in the delivered power-to-size ratio of semiconductor lasers for compact implementation. We present a new two-capacitor memory-cell configuration that provides an enhanced bipolar optoelectronic response from the edges of the capacitor at incident power as low as ~ 2 mW/μm(2). A novel device design based on lead zirconate titanate with the c axis parallel to the substrate is suggested to reduce the requirement of incident optical power further by orders of magnitude.     

ZnxCd1−xS/Cu2S heterojunction solar cells—I: Fabrication and performance

Alloy films of Zn_xCd_1?xS have been prepared by both vacuum evaporation and spray pyrolysis. Junctions of these films with Cu₂S have been formed by the solid state reaction for the evaporated alloy films and by the conventional chemiplating technique for the spray deposited alloy films. The current-voltage characteristics, diffusion length of minority carriers and spectral response of the cells have been determined. Highest conversion efficiency for the evaporated cells is 6.5% corresponding to x = 0 and for the sprayed cells is 5.6% corresponding to x = 0.1. The cell performance has been analysed and further scope of improvement indicated.     

Optically addressed ferroelectric memory and its applications

Abstract

Thin films of perovskite titanates can be composition-tailored to exhibit ferroelectric, pyroelectric, or piezoelectric properties, in varying degree of combinations, and thereby modulating their response when illuminated with light. This paper reviews the potential applications of photoresponse from lead zirconate titanate thin films. In ferroelectric materials, such as lead zirconate titanate (PZT), for example, the photoresponse shows a clear dependence on remanent polarization. The main highlight of the paper is a review of the concept of optically addressed ferroelectric memory and identification of its high impact applications. Incidence of energetic laser beam pulses gives rise to two different classes of phenomena: first, a thermally triggered piezoelectric/pyroelectric response and second, an optoelectronic response. Optimizing the device geometry and selecting the illumination characteristics, one can easily control the dominating mechanism in a device. The optoelectronic effect emerges at a relatively lower incident beam power (≤2 mW/μm²) and can be as fast as the duration of the pulse (≤10ns), and therefore, is directly relevant as a non-destructive read-out signal from optically addressed ferroelectric memory. The optical NDRO signal offers itself as a unique tool that allows a non-destructive ‘probe’ for the capacitor, without causing any polarization switching in it, the characteristic artifact of the electrical destructive readout (DRO) measurement technique. Furthermore, the polarization dependent photoresponse could also be exploited as a non-destructive evaluation tool for mapping the domains in ferroelectric films, to generate fundamental understanding of the domain dynamics (generation, growth, movement, etc., under applied field, illumination, and/or temperature) in such thin films. In addition, with a high fidelity analog nature of the remanent polarization, the highly parallel, high speed photoresponse output from such optically addressed memories may be ideally suited for high performance computing applications especially involving image processing, high speed communication and parallel processing with architectures, such as large scale artificial neural networks.     

Analog VLSI neural networks: implementation issues and examples inoptimization and supervised learning

Time-critical neural network applications that require fully parallel hardware implementations for maximal throughput are considered. The rich array of technologies that are being pursued is surveyed, and the analog CMOS VLSI medium approach is focused on. This medium is messy in that limited dynamic range, offset voltages, and noise sources all reduce precision. The authors examine how neural networks can be directly implemented in analog VLSI, giving examples of approaches that have been pursued to date. Two important application areas are highlighted: optimization, because neural hardware may offer a speed advantage of orders of magnitude over other methods; and supervised learning, because of the widespread use and generality of gradient-descent learning algorithms as applied to feedforward networks     

Metal chalcogenide-oxide composite coatings prepared by spray pyrolysis

The spray pyrolysis technique has been employed to deposit composite coatings of chalcogenides of cadmium, zinc, lead and cobalt with oxides of aluminium, tin, lead, zinc and cobalt. Widely varying microstructural, electronic, optical and chemical properties have been obtained for such layers by monitoring the oxide composition, its spatial distribution and profile along the thickness. The large area chalcogenide-oxide composite films prepared by this technique are eminently suited for photovoltaic energy conversion, photothermal energy conversion and voltage-dependent resistor (Varistor) applications.In this paper we report our studies on co-pyrolytically deposited CdS:Al₂O₃ and CdS:SnO₂ layers and their application to improved thin film solar cells. Each of the oxides is insoluble in CdS and is segregated at the grain boundaries in the deposited films. Small amounts (less than 10%) of oxide in CdS are found to reduce its grain size negligibly and to make the film more compact, hard, adherent and less susceptible to chemical attack. The altered microstructure modifies the surface topography of the CdS film from a pebble-like roughness to an improved void-free serpentine texture. Segregated oxide in CdS does not affect the optical band gap of the films, although the composites exhibit enhanced diffuse optical scattering.Large area CdS films with a gradient profile of oxide have been utilized to fabricate thin film CdS/Cu₂S solar cells. The growth (length and distribution) of Cu₂S fingers and/or curtains deep into the top CdS layers during the topotaxial conversion reaction of chemiplating is controlled by the presence of oxide along the grain boundaries. This has not only resulted in improved interface topography for better carrier collection and reduced shunt losses but has also enabled us to decrease drastically the CdS film thickness necessary for the solar cells. Furthermore, the subsequent degradation of the junction via the well-known mechanism of the loss of copper from the Cu₂S layer by diffusion into CdS is expected to be considerably reduced by the presence of the oxide gradient in the CdS layer.     

Review: The Benefits and Applications of Bioinspired Flight Capabilities

This paper addresses the challenges of flight on Mars that at this time have the same element of novelty as flight on Earth itself was a novelty in the Kitty Hawk era almost 100 years ago, details the scientific need for such flyers, highlights the bioinspired engineering of exploration systems (BEES) flyer development and finally describes a few viable mission architecture options that allow reliable data return from the BEES flyers using the limited telecom infrastructure that can be made available with a lander base to orbiter combination on Mars. Our recent developments using inspiration from biology that are enabling the pathway to demonstrate flight capability for Mars exploration are described. These developments hold substantial spin‐offs for a variety of applications both for NASA and DoD. Unmanned exploration to date suggests that Mars once had abundant liquid water (considered essential for life as we know it). It is not clear what transpired on the Martian climate to have turned the planet into the desert that it is today. Developing a comprehensive understanding of the past and present climatic events for Mars may provide important information relevant to the future of our own planet. Such exploration missions are enabled using the BEES technology. © 2003 Wiley Periodicals, Inc.