Engineering topics

Ahumada, Albert J., Jr.

Publications and source records attributed to Ahumada, Albert J., Jr..

35 records · Page 2

NASA NTRS1993

A visual detection model for DCT coefficient quantization

The discrete cosine transform (DCT) is widely used in image compression, and is part of the JPEG and MPEG compression standards. The degree of compression, and the amount of distortion in the decompressed image are determined by the quantization of the transform coefficients. The standards do not specify how the DCT coefficients should be quantized. Our approach is to set the quantization level for each coefficient so that the quantization error is at the threshold of visibility. Here we combine results from our previous work to form our current best detection model for DCT coefficient quantization noise. This model predicts sensitivity as a function of display parameters, enabling quantization matrices to be designed for display situations varying in luminance, veiling light, and spatial frequency related conditions (pixel size, viewing distance, and aspect ratio). It also allows arbitrary color space directions for the representation of color.

Ahumada, Albert J., Jr.↗

NASA NTRS1992

Color matrix display simulation based upon luminance and chromatic contrast sensitivity of early vision

This paper describes the design and operation of a new simulation model for color matrix display development. It models the physical structure, the signal processing, and the visual perception of static displays, to allow optimization of display design parameters through image quality measures. The model is simple, implemented in the Mathematica computer language, and highly modular. Signal processing modules operate on the original image. The hardware modules describe backlights and filters, the pixel shape, and the tiling of the pixels over the display. Small regions of the displayed image can be visualized on a CRT. Visual perception modules assume static foveal images. The image is converted into cone catches and then into luminance, red-green, and blue-yellow images. A Haar transform pyramid separates the three images into spatial frequency and direction-specific channels. The channels are scaled by weights taken from human contrast sensitivity measurements of chromatic and luminance mechanisms at similar frequencies and orientations. Each channel provides a detectability measure. These measures allow the comparison of images displayed on prospective devices and, by that, the optimization of display designs.

Martin, Russel A.↗

NASA NTRS1992

Principled halftoning based on human vision models

When models of human vision adequately measure the relative quality of candidate halftonings of an image, the problem of halftoning the image becomes equivalent to the search problem of finding a halftone that optimizes the quality metric. Because of the vast number of possible halftones, and the complexity of image quality measures, this principled approach has usually been put aside in favor of fast algorithms that seem to perform well. We find that the principled approach can lead to a range of useful halftoning algorithms, as we trade off speed for quality by varying the complexity of the quality measure and the thoroughness of the search. High quality halftones can be obtained reasonably quickly, for example, by using as a measure the vector length of the error image filtered by a contrast sensitivity function, and, as the search procedure, the sequential adjustment of individual pixels to improve the quality measure. If computational resources permit, simulated annealing can find nearly optimal solutions.

Mulligan, Jeffrey B.↗

NASA NTRS1992

Luminance-model-based DCT quantization for color image compression

A model is developed to approximate visibility thresholds for discrete cosine transform (DCT) coefficient quantization error based on the peak-to-peak luminance of the error image. Experimentally measured visibility thresholds for R, G, and B DCT basis functions can be predicted by a simple luminance-based detection model. This model allows DCT coefficient quantization matrices to be designed for display conditions other than those of the experimental measurements: other display luminances, other veiling luminances, and other spatial frequencies (different pixel spacings, viewing distances, and aspect ratios).

Ahumada, Albert J., Jr.↗

NASA NTRS1991

Network compensation for missing sensors

A network learning translation invariance algorithm to compute interpolation functions is presented. This algorithm with one fixed receptive field can construct a linear transformation compensating for gain changes, sensor position jitter, and sensor loss when there are enough remaining sensors to adequately sample the input images. However, when the images are undersampled and complete compensation is not possible, the algorithm need to be modified. For moderate sensor losses, the algorithm works if the transformation weight adjustment is restricted to the weights to output units affected by the loss.

Ahumada, Albert J., Jr.↗

NASA NTRS1990

Pyramidal Image-Processing Code For Hexagonal Grid

Algorithm based on processing of information on intensities of picture elements arranged in regular hexagonal grid. Called "image pyramid" because image information at each processing level arranged in hexagonal grid having one-seventh number of picture elements of next lower processing level, each picture element derived from hexagonal set of seven nearest-neighbor picture elements in next lower level. At lowest level, fine-resolution of elements of original image. Designed to have some properties of image-coding scheme of primate visual cortex.

Watson, Andrew B.↗

NASA NTRS1990

Sampling and noise in vision networks

This research is part of the Human Interface Research Branch-Vision Group's program to develop computable models of biological solutions to general vision system problems. Two problem areas are addressed: effects of discrete sampling by receptors, and effects of visual system noise.

Ahumada, Albert J., Jr.↗

NASA NTRS1990

Networks for image acquisition, processing and display

The human visual system comprises layers of networks which sample, process, and code images. Understanding these networks is a valuable means of understanding human vision and of designing autonomous vision systems based on network processing. Ames Research Center has an ongoing program to develop computational models of such networks. The models predict human performance in detection of targets and in discrimination of displayed information. In addition, the models are artificial vision systems sharing properties with biological vision that has been tuned by evolution for high performance. Properties include variable density sampling, noise immunity, multi-resolution coding, and fault-tolerance. The research stresses analysis of noise in visual networks, including sampling, photon, and processing unit noises. Specific accomplishments include: models of sampling array growth with variable density and irregularity comparable to that of the retinal cone mosaic; noise models of networks with signal-dependent and independent noise; models of network connection development for preserving spatial registration and interpolation; multi-resolution encoding models based on hexagonal arrays (HOP transform); and mathematical procedures for simplifying analysis of large networks.

Ahumada, Albert J., Jr.↗

NASA NTRS1990

Learning receptor positions from imperfectly known motions

An algorithm is described for learning image interpolation functions for sensor arrays whose sensor positions are somewhat disordered. The learning is based on failures of translation invariance, so it does not require knowledge of the images being presented to the visual system. Previously reported implementations of the method assumed the visual system to have precise knowledge of the translations. It is demonstrated that translation estimates computed from the imperfectly interpolated images can have enough accuracy to allow the learning process to converge to a correct interpolation.

Ahumada, Albert J., Jr.↗

NASA NTRS1989

A hexagonal orthogonal-oriented pyramid as a model of image representation in visual cortex

Retinal ganglion cells represent the visual image with a spatial code, in which each cell conveys information about a small region in the image. In contrast, cells of the primary visual cortex use a hybrid space-frequency code in which each cell conveys information about a region that is local in space, spatial frequency, and orientation. A mathematical model for this transformation is described. The hexagonal orthogonal-oriented quadrature pyramid (HOP) transform, which operates on a hexagonal input lattice, uses basis functions that are orthogonal, self-similar, and localized in space, spatial frequency, orientation, and phase. The basis functions, which are generated from seven basic types through a recursive process, form an image code of the pyramid type. The seven basis functions, six bandpass and one low-pass, occupy a point and a hexagon of six nearest neighbors on a hexagonal lattice. The six bandpass basis functions consist of three with even symmetry, and three with odd symmetry. At the lowest level, the inputs are image samples. At each higher level, the input lattice is provided by the low-pass coefficients computed at the previous level. At each level, the output is subsampled in such a way as to yield a new hexagonal lattice with a spacing square root of 7 larger than the previous level, so that the number of coefficients is reduced by a factor of seven at each level. In the biological model, the input lattice is the retinal ganglion cell array. The resulting scheme provides a compact, efficient code of the image and generates receptive fields that resemble those of the primary visual cortex.

Watson, Andrew B.↗

NASA NTRS1989

Reconstructing irregularly sampled images by neural networks

Neural-network-like models of receptor position learning and interpolation function learning are being developed as models of how the human nervous system might handle the problems of keeping track of the receptor positions and interpolating the image between receptors. These models may also be of interest to designers of image processing systems desiring the advantages of a retina-like image sampling array.

Ahumada, Albert J., Jr.↗

NASA NTRS1988

Simulation models of early visual processes

Several areas of early visual processes are studied using computer models. These models include retinal cone placement, cone color arrangement, development of geniculate receptive fields, cortical simple cells, and motion field extraction. The receptive field of a model cortical unit is indicated schematically.

Ahumada, Albert J., Jr.↗

NASA NTRS1987

An orthogonal oriented quadrature hexagonal image pyramid

An image pyramid has been developed with basis functions that are orthogonal, self-similar, and localized in space, spatial frequency, orientation, and phase. The pyramid operates on a hexagonal sample lattice. The set of seven basis functions consist of three even high-pass kernels, three odd high-pass kernels, and one low-pass kernel. The three even kernels are identified when rotated by 60 or 120 deg, and likewise for the odd. The seven basis functions occupy a point and a hexagon of six nearest neighbors on a hexagonal sample lattice. At the lowest level of the pyramid, the input lattice is the image sample lattice. At each higher level, the input lattice is provided by the low-pass coefficients computed at the previous level. At each level, the output is subsampled in such a way as to yield a new hexagonal lattice with a spacing sq rt 7 larger than the previous level, so that the number of coefficients is reduced by a factor of 7 at each level. The relationship between this image code and the processing architecture of the primate visual cortex is discussed.

Watson, Andrew B.↗

NASA NTRS1987

Putting the visual system noise back in the picture

Computable expressions for the input-picture-equivalent contrast noise of the visual system are provided for the locally linear subclass of nonlinear models, where the internal model noise is allowed to be signal dependent. The equivalent-noise concept is thereby extended to many of the models developed to explain masking and discrimination among suprathreshold stimuli. For these models the equivalent noise depends on the masking stimulus, and its structure can be strongly determined by the representation of the masker at the level of the system at which the performance-limiting noise is generated. The expressions are applicable to the case of less-than-full-rank transformations. Pictures that have hypothetical visual-system noise projected back into them can provide insights into efficient picture-coding algorithms.

Ahumada, Albert J., Jr.↗

NASA NTRS1987

Cone sampling array models

A model is described for positioning cones in the retina. Each cone has a circular disk of influence, and the disks are tightly packed outward from the center. This model has three parameters that can vary with eccentricity: the mean radius of the cone disk, the standard deviation of the cone disk radius, and the standard deviation of postpacking jitter. Estimates for these parameters out to 1.6 deg are found by using measurements reported by Hirsch and Hylton (1985) and Hirsch and Miller (1987) of the positions of the cone inner segments of an adult macaque. The estimation is based on fitting measures of variation in local intercone distances, and the fit to these measures is good.

Ahumada, Albert J., Jr.↗

NASA NTRS1987

Simulated annealing in networks for computing possible arrangements for red and green cones

Attention is given to network models in which each of the cones of the retina is given a provisional color at random, and then the cones are allowed to determine the colors of their neighbors through an iterative process. A symmetric-structure spin-glass model has allowed arrays to be generated from completely random arrangements of red and green to arrays with approximately as much disorder as the parafoveal cones. Simulated annealing has also been added to the process in an attempt to generate color arrangements with greater regularity and hence more revealing moirepatterns than than the arrangements yielded by quenched spin-glass processes. Attention is given to the perceptual implications of these results.

Ahumada, Albert J., Jr.↗

NASA NTRS1986

Use of a raster framebuffer in vision research

For the study of human and animal vision, the video framebuffer is the only technology that is capable of displaying two-dimensional images with precise control of contrast, luminance, and display timing. The video framebuffer also allows precise control of color. However, this device is not designed for precise psychophysical displays, and techniques must be developed to use them in this role. In order to be concrete, an RDS-3000 raster display system (Adage, 1982) is used which is hosted by a PDP-11/73 under the Venix operating system. The principles generalize to other machines. Where it clarifies the issues, the syntax of the software routines involved is shown.

Watson, Andrew B.↗