Sally L. Wood

Language identification for printed text independent of segmentation

This paper presents efficient algorithms for determining the language classification of machine generated documents without requiring the identification of individual characters. Such algorithms may be useful for sorting and routing of facsimile documents as they arrive so that appropriate routing and secondary analysis, which may include OCR, is selected for each document. It may also prove useful as a component of a content addressable document access system. There have been numerous reported efforts which attempt to segment printed documents into homogeneous regions using Hough transforms, hidden Markov models, morphological filtering, and neural networks. However, language identification can be accomplished without explicit segmentation using the less computationally intensive methods described.

PURSUING AUTOMATED CLASSIFICATION OF HISTORIC PHOTOGRAPHIC PAPERS FROM RAKING LIGHT IMAGES

Abstract Surface texture is a critical feature in the manufacture, marketing, and use of photographic paper. Raking light reveals texture through a stark rendering of highlights and shadows. Though close-up raking light images effectively document surface features of photographic paper, the sheer number and diversity of textures used for historic papers prohibits efficient visual classification. This work provides evidence that automatic, computer-based classification of texture documented with raking light is feasible by demonstrating an encouraging degree of success sorting a set of 120 images made from samples of historic silver gelatin paper. Using this dataset, four university teams applied different image-processing strategies for automatic feature extraction and degree of similarity quantification. All four approaches successfully detected strong affinities and outliers built into the dataset. The creation and deployment of the algorithms was carried out by the teams without prior knowledge of the distributions of similarities and outliers. These results indicate that automatic classification of silver gelatin photographic paper based on close-up texture images is feasible and should be pursued. To encourage the development of other classification schemes, the 120-sample “training” dataset used in this work is available to other academic researchers at http://www.PaperTextureID.org.

Adaptive flat multiresolution multiplexed computational imaging architecture utilizing micromirror arrays to steer subimager fields of view

A thin, agile multiresolution, computational imaging sensor architecture, termed PANOPTES (processing arrays of Nyguist-limited observations to produce a thin electro-optic sensor), which utilizes arrays of microelectromechanical mirrors to adaptively redirect the fields of view of multiple low-resolution subimagers, is described. An information theory-based algorithm adapts the system and restores the image. The modulation transfer function (MTF) effects of utilizing micromirror arrays to steering imaging systems are analyzed, and computational methods for combining data collected from systems with differing MTFs are presented.

High spatial resolution single multiwalled carbon nanotube electrode for stimulation, recording, and whole cell voltage clamping of electrically active cells

We report the stimulation, recording, and voltage clamp of muscle fibers using a 30 nm diameter single multiwalled carbon nanotube electrode (sMWNT electrode) tip. Because of the lower access resistance, the sMWNT electrode conducts extracellular and intracellular stimulation more efficiently compared to glass micropipettes. The sMWNT electrode records field potentials and action potentials and performs whole cell voltage clamping of single fibers.

Resolution improvement for compound eye images through lens diversity

Reconstruction of high resolution images from multiple low resolution images at various displacements is a well studied, ill posed problem. Designs using lenses with different imaging characteristics improve the theoretical results and also reduce the image reconstruction problem to a set of loosely coupled smaller reconstructions. The paper derives the performance limits for reconstruction from multiple lower resolution images as a function of measurement bit precision and measurement noise.

Modem constellation identification: a performance comparison of two methods

Two algorithms are described for determining the signal constellation quadrature amplitude modulation in use on a (QAM) modulator, one based on Radon transform ideas and the other based on histogramming received signal radius information. The detection performances of the two specific algorithms are compared as a function of the signal-to-disturbance power ratio and the number of samples available. Generally speaking, it is shown that, when the carrier frequency is known with reasonable accuracy, the Radon transform method performs much better than the specific radius- or magnitude-only method tested, in the sense that for the same signal-to-disturbance ratio the former can reliably determine the constellation size and orientation with about one-tenth of the data samples required by the latter. It should be noted, however, that neither of the two algorithms has been optimized, nor have the comparisons taken into account other issues such as the details of particular applications, the relative computational requirements, and the added complexity needed by the Radon transform method to deal with uncertainties in the carrier frequency.<<ETX>>

Understanding the topology of a telephone network via internally-sensed network tomography

The ability to determine the topology of worldwide telephone networks offers the promise of substantially improving their operating efficiency. This paper explores the problem of identifying the topology of a telephone network using observations made within the network. Using tomographic methods inspired by medical imaging, we consider measurements made by transmitting probes (e.g., phone calls) between network endpoints. In general, these measurements alone do not suffice to reconstruct a unique network, and in fact, there are many network topologies from which the set of measurements could have been generated. We propose a topology reconstruction algorithm based on correlating measurements collected at different internal nodes, and identify conditions under which correctness of the inferred topology is guaranteed.

Convergence rate limitations in certain frequency-domain adaptive filters

It is shown that the transport delay present in the feedback path of a frequency-domain adaptive filter consisting of the combination of a CMA (constant modulus algorithm) with a transmultiplexer reduces the filters maximum attainable convergence rate. An upper bound for the convergence rate is developed in terms of the amount of group delay and hardware pipeline delay present in the adaptive system under analysis. These limitations are shown to be fundamental in nature. They relate directly to the amount of spectral resolution desired in the adaptive filter and not to the particular nonlinear error function employed.<<ETX>>

Benefits of optical system diversity for multiplexed image reconstruction

Algorithms that use optical system diversity to improve multiplexed image reconstruction from multiple low-resolution images are analyzed and demonstrated. Compared with systems using identical imagers, systems using additional lower-resolution imagers can have improved accuracy and computation. The diverse system is not sensitive to boundary conditions and can take full advantage of improvements that decrease noise and allow an increased number of bits per pixel to represent spatial information in a scene.

Empirical study of unipolar and bipolar configurations using high resolution single multi-walled carbon nanotube electrodes for electrophysiological probing of electrically excitable cells.

Identifying the neurophysiological basis underlying learning and memory in the mammalian central nervous system requires the development of biocompatible, high resolution, low electrode impedance electrophysiological probes; however, physically, electrode impedance will always be finite and, at times, large. Herein, we demonstrate through experiments performed on frog sartorius muscle that single multi-walled carbon nanotube electrode (sMWNT electrode) geometry and placement are two degrees of freedom that can improve biocompatibility of the probe and counteract the detrimental effects of MWNT/electrolyte interface impedance on the stimulation efficiency and signal-to-noise ratio (SNR). We show that high aspect ratio dependent electric field enhancement at the MWNT tip can boost stimulation efficiency. Derivation of the sMWNT electrodes electrical equivalent indicates that, at low stimulus voltage regimes below 1 V, current conduction is mediated by charge fluctuation in the double layer obviating electrolysis of water, which is potentially toxic to pH sensitive biological tissue. Despite the accompanying increase in electrode impedance, a pair of closely spaced sMWNT electrodes in a two probe (bipolar) configuration maintains biocompatibility and enhances stimulation efficiency and SNR compared to the single probe (unipolar) configuration. For stimulus voltages below 1 V, the electrical equivalent verifies that current conduction in the two probe configuration still proceeds via charge fluctuation in the double layer. As an extracellular stimulation electrode, the two sMWNT electrodes comprise a current dipole that concentrates the electric field and the current density in a smaller region of sartorius; consequently, the bipolar configuration can elicit muscle fiber twitching at low voltages that preclude electrolysis of water. When recording field potentials, the bipolar configuration subtracts the potential between two points allowing for the detection of higher signal amplitudes. As a result, SNR is improved. These results indicate that use of the high aspect ratio MWNT in a bipolar configuration can achieve a biocompatible electrode that offers enhanced stimulation efficiency and higher SNR.