Richard W. Christiansen

A shadow detection and removal algorithm for 2-D images

An algorithm that detects and removes background shadows from images in which the pattern set occupies the upper-most intensity range of the image and the image is background dominant outside the pattern set is presented. The algorithm will remove background shadows and preserve any remaining texture left behind by the shadow function. A mathematical model of the histogram modification function of the shadow-removal algorithm is developed. An analysis of the sequential nature of the algorithm is included along with simulated results to verify the mathematical model developed and to show the effectiveness of the algorithm in removing background pattern shadows.<<ETX>>

Application of a human auditory model to loudness perception and hearing compensation

A model is proposed which mathematically transforms an acoustic stimulus into a form which is believed to be more nearly related to that used by the auditory cortex to interpret the sound. The model is based upon research towards understanding the response of the human auditory system to sound stimuli. The motivation and approach in developing this model follow the philosophy pursued in the development of a similar model to the problem of hearing compensation for impaired individuals is shown to yield a bank of bandpass filters each followed by a homomorphic multiplicative AGC. Clinical tests on hearing impaired subjects suggest that this approach is far superior to other current hearing compensation schemes.

Multispectral document enhancement: ancient carbonized scrolls

Multispectral imaging of carbonized documents has been shown to be an effective diagnostic tool as well as providing for spectral analysis and enhancement. The Herculaneum scroll diagnostic data showed that sufficient contrast could be obtained with an image at a single wavelength in the near infrared. On the other hand, the Petra scroll multispectral data demonstrates the analytical and enhancement capabilities provided by this form of document conservation and archiving.

Multispectral analysis of ancient Maya pigments: implications for the Naj Tunich corpus

Naj Tunich, in southeastern Peten, Guatemala, is one of only five caves currently known to contain Maya hieroglyphic writing. It is universally recognized as the preeminent Maya cave site because its corpus ofinscriptions exceeds those of the other four caves combined. In June of 1998, over half of these inscriptions were documented at visible and near-infrared wavelengths using multi-spectral imaging techniques. Spectral differences were noted in the Naj Tunich images especially in the infrared. Spectral signatures were used to identify differences in the Naj Tunich pigments and suggest that at least three different pigments were used. More importantly, the ability to document spectral differences reveals far more complexity in the Naj Tunich corpus than previously appreciated. Several instances of over-painting, repainting or touching up were discovered. The cases of over-painting reveal the temporal complexity of the drawings and suggest that the painting occurred over a longer period of time than had been proposed by previous investigation. The ability to characterize pigment composition using spectral data has also led the authors to question a number of relational differences and similarities between drawings proposed on the basis of stylistic analysis.

Detection algorithms for image sequence analysis

The development of five single-stage and four multistage algorithms is reported. A comparison of the performance, for moving target identification applications, of 13 algorithms on a simulated infrared (IR) image database is included. Several of the algorithms show significant improvement in performance over the well-known algorithms in the area. >

Quality measures for SAR images

Synthetic aperture radar (SAR) contains information features different from those of infrared images and conventional photography. Image quality measures specifically created for SAR are necessary to measure SAR processor performance especially in the context of lossy image data compression. The quality measures discussed include several existing measures and newly developed measures created to better gauge specific features of SAR imagery. Each quality performance measure is evaluated subjectively to determine the applicability to SAR image fidelity.<<ETX>>

Image Processing For Target Detection Using Data From A Staring Mosaic Infrared Sensor In Geosynchronous Orbit

A set of multistage image processing algorithms has been developed to perform change detection in image sequence analysis. These algorithms are tailored to the task of moving target identification (MTI). Conventional differencing algorithms are compared to a new spatial differencing algorithm. Subsequent processing is accomplished with thresholding and a suboptimal implementation of a 2-D matched filter designed for simplicity of calculation. The image processing algorithms are then coupled with a simple tracking/detection algorithm. The resulting combination of processing shows very good performance for detecting targets in simulated mosaic IR images, with a projected probability of false alarm less than one per 1011 frames and a probability of detection for targets within the model approaching unity.

Noise reduction in speech using adaptive filtering I: Signal processing algorithms

Adaptive filtering is employed in configurations to filter narrow‐band speech corrupted by noise. These configurations utilize either an independent sample of the input noise or rely on correlation properties of the speech to accomplish cancellation. Necessary constraints on the algorithms to retain and/or improve intelligibility for normal and hearing‐impaired populations are presented. Previous work with the LMS algorithm is compared to the performance resulting from new methods which address the problem created by widespread eigenvalues in the speech plus noise correlation matrix. It is shown that failure to deal with this problem in the single reference application results in a muffling resonant distortion of the speech spectrum. Spectral averages of the original and processed versions of the speech segments are compared to determine improvement. Auditory tests are presented in the companion paper.

Application of the LMS adaptive filter to improve speech communication in the presence of noise

A major complaint of individuals with normal hearing and hearing impairment is a reduced ability to understand speech in a noisy environment. An adaptive LMS filter was employed to process speech in signal-to-noise ratios (S/N) varying from -8 to +12 dB. The filter configuration is commonly called noise cancellation. Intelligibility measures were obtained on normal and hearing impaired individuals using CID-W22 word lists. For the normal subjects mean intelligibility scores unprocessed were near zero for the S/N ratios -8, -4, and 0 dB and increased from 30 to 50% for the same S/N ratios when processed. For S/N ratios of +4 through +12, mean intelligibility scores prior to filtering ranged from 24 to 46%; after filtering, the scores improved from 67 to 75% i.e., within 7% of the nolse-free sound-field recordings. Preliminary data for the hearing impaired showed the same pattern with their improvement also approaching the scores obtained with no noise present in the sound field.

A small, manned aircraft as a testbed for radar sensor development

The flight testing phase is vital in the development of an airborne SAR system, but can be time consuming and expensive, especially for UAS based systems. As part of a SAR design methodology, we are using a small, manned aircraft as a surrogate for UASs and other platforms. Prototypes of new systems can be easily installed on the testbed in order to quickly and inexpensively obtain sensor and motion data. These data can be used to aid in system-specific algorithm development, as well as further refinement of the system hardware as necessary.