Engineering topics

Watson, Andrew B.

Publications and source records attributed to Watson, Andrew B..

At least 55 records · Page 3

NASA NTRS1994

Image Discrimination Models Predict Object Detection in Natural Backgrounds

Object detection involves looking for one of a large set of object sub-images in a large set of background images. Image discrimination models only predict the probability that an observer will detect a difference between two images. In a recent study based on only six different images, we found that discrimination models can predict the relative detectability of objects in those images, suggesting that these simpler models may be useful in some object detection applications. Here we replicate this result using a new, larger set of images. Fifteen images of a vehicle in an other-wise natural setting were altered to remove the vehicle and mixed with the original image in a proportion chosen to make the target neither perfectly recognizable nor unrecognizable. The target was also rotated about a vertical axis through its center and mixed with the background. Sixteen observers rated these 30 target images and the 15 background-only images for the presence of a vehicle. The likelihoods of the observer responses were computed from a Thurstone scaling model with the assumption that the detectabilities are proportional to the predictions of an image discrimination model. Three image discrimination models were used: a cortex transform model, a single channel model with a contrast sensitivity function filter, and the Root-Mean-Square (RMS) difference of the digital target and background-only images. As in the previous study, the cortex transform model performed best; the RMS difference predictor was second best; and last, but still a reasonable predictor, was the single channel model. Image discrimination models can predict the relative detectabilities of objects in natural backgrounds.

Ahumada, Albert J., Jr.↗

NASA NTRS1993

DCT quantization matrices visually optimized for individual images

This presentation describes how a vision model incorporating contrast sensitivity, contrast masking, and light adaptation is used to design visually optimal quantization matrices for Discrete Cosine Transform image compression. The Discrete Cosine Transform (DCT) underlies several image compression standards (JPEG, MPEG, H.261). The DCT is applied to 8x8 pixel blocks, and the resulting coefficients are quantized by division and rounding. The 8x8 'quantization matrix' of divisors determines the visual quality of the reconstructed image; the design of this matrix is left to the user. Since each DCT coefficient corresponds to a particular spatial frequency in a particular image region, each quantization error consists of a local increment or decrement in a particular frequency. After adjustments for contrast sensitivity, local light adaptation, and local contrast masking, this coefficient error can be converted to a just-noticeable-difference (jnd). The jnd's for different frequencies and image blocks can be pooled to yield a global perceptual error metric. With this metric, we can compute for each image the quantization matrix that minimizes the bit-rate for a given perceptual error, or perceptual error for a given bit-rate. Implementation of this system demonstrates its advantages over existing techniques. A unique feature of this scheme is that the quantization matrix is optimized for each individual image. This is compatible with the JPEG standard, which requires transmission of the quantization matrix.

Watson, Andrew B.↗

NASA NTRS1993

Visual optimization of DCT quantization matrices for individual images

Many image compression standards (JPEG, MPEG, H.261) are based on the Discrete Cosine Transform (DCT). However, these standards do not specify the actual DCT quantization matrix. We have previously provided mathematical formulae to compute a perceptually lossless quantization matrix. Here I show how to compute a matrix that is optimized for a particular image. The method treats each DCT coefficient as an approximation to the local response of a visual 'channel'. For a given quantization matrix, the DCT quantization errors are adjusted by contrast sensitivity, light adaptation, and contrast masking, and are pooled non-linearly over the blocks of the image. This yields an 8x8 'perceptual error matrix'. A second non-linear pooling over the perceptual error matrix yields total perceptual error. With this model we may estimate the quantization matrix for a particular image that yields minimum bit rate for a given total perceptual error, or minimum perceptual error for a given bit rate. Custom matrices for a number of images show clear improvement over image-independent matrices. Custom matrices are compatible with the JPEG standard, which requires transmission of the quantization matrix.

Watson, Andrew B.↗

NASA NTRS1992

Separability of spatiotemporal spectra of image sequences

The spatiotemporal power spectrum was calculated of 14 image sequences in order to determine the degree to which the spectra are separable in space and time, and to assess the validity of the commonly used exponential correlation model found in the literature. The spectrum was expanded by a Singular Value Decomposition into a sum of separable terms and an index was defined of spatiotemporal separability as the fraction of the signal energy that can be represented by the first (largest) separable term. All spectra were found to be highly separable with an index of separability above 0.98. The power spectra of the sequences were well fit by a separable model. The power spectrum model corresponds to a product of exponential autocorrelation functions separable in space and time.

Eckert, Michael P.↗

NASA NTRS1992

Supercomputers Of The Future

Report evaluates supercomputer needs of five key disciplines: turbulence physics, aerodynamics, aerothermodynamics, chemistry, and mathematical modeling of human vision. Predicts these fields will require computer speed greater than 10(Sup 18) floating-point operations per second (FLOP's) and memory capacity greater than 10(Sup 15) words. Also, new parallel computer architectures and new structured numerical methods will make necessary speed and capacity available.

Peterson, Victor L.↗

NASA NTRS1992

Transfer of contrast sensitivity in linear visual networks

Contrast sensitivity is a useful measure of the ability of an observer to distinguish contrast signals from noise. Although usually applied to human observers, contrast sensitivity can also be defined operationally for individual visual neurons. In a model linear neuron consisting of a filter and noise source, this operational measure is a function of filter gain, noise power spectrum, signal duration, and a performance criterion. This definition allows one to relate the sensitivities of linear neurons at different levels in the visual pathway. Mathematical formulas describing these relationships are derived, and the general model is applied to the specific problem of relating the sensitivities of parvocellular LGN neurons and cortical simple cells in the primate.

Watson, Andrew B.↗

NASA NTRS1991

Digital visual communications using a Perceptual Components Architecture

The next era of space exploration will generate extraordinary volumes of image data, and management of this image data is beyond current technical capabilities. We propose a strategy for coding visual information that exploits the known properties of early human vision. This Perceptual Components Architecture codes images and image sequences in terms of discrete samples from limited bands of color, spatial frequency, orientation, and temporal frequency. This spatiotemporal pyramid offers efficiency (low bit rate), variable resolution, device independence, error-tolerance, and extensibility.

Watson, Andrew B.↗

NASA NTRS1990

Perceptual-components architecture for digital video

A perceptual-components architecture for digital video partitions the image stream into signal components in a manner analogous to that used in the human visual system. These components consist of achromatic and opponent color channels, divided into static and motion channels, further divided into bands of particular spatial frequency and orientation. Bits are allocated to an individual band in accord with visual sensitivity to that band and in accord with the properties of visual masking. This architecture is argued to have desirable features such as efficiency, error tolerance, scalability, device independence, and extensibility.

Watson, Andrew B.↗

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

Vision Science and Technology at NASA: Results of a Workshop

A broad review is given of vision science and technology within NASA. The subject is defined and its applications in both NASA and the nation at large are noted. A survey of current NASA efforts is given, noting strengths and weaknesses of the NASA program.

Watson, Andrew B.↗

NASA NTRS1990

Ames vision group research overview

A major goal of the reseach group is to develop mathematical and computational models of early human vision. These models are valuable in the prediction of human performance, in the design of visual coding schemes and displays, and in robotic vision. To date researchers have models of retinal sampling, spatial processing in visual cortex, contrast sensitivity, and motion processing. Based on their models of early human vision, researchers developed several schemes for efficient coding and compression of monochrome and color images. These are pyramid schemes that decompose the image into features that vary in location, size, orientation, and phase. To determine the perceptual fidelity of these codes, researchers developed novel human testing methods that have received considerable attention in the research community. Researchers constructed models of human visual motion processing based on physiological and psychophysical data, and have tested these models through simulation and human experiments. They also explored the application of these biological algorithms to applications in automated guidance of rotorcraft and autonomous landing of spacecraft. Researchers developed networks for inhomogeneous image sampling, for pyramid coding of images, for automatic geometrical correction of disordered samples, and for removal of motion artifacts from unstable cameras.

Watson, Andrew B.↗

NASA NTRS1990

Pyramid image codes

All vision systems, both human and machine, transform the spatial image into a coded representation. Particular codes may be optimized for efficiency or to extract useful image features. Researchers explored image codes based on primary visual cortex in man and other primates. Understanding these codes will advance the art in image coding, autonomous vision, and computational human factors. In cortex, imagery is coded by features that vary in size, orientation, and position. Researchers have devised a mathematical model of this transformation, called the Hexagonal oriented Orthogonal quadrature Pyramid (HOP). In a pyramid code, features are segregated by size into layers, with fewer features in the layers devoted to large features. Pyramid schemes provide scale invariance, and are useful for coarse-to-fine searching and for progressive transmission of images. The HOP Pyramid is novel in three respects: (1) it uses a hexagonal pixel lattice, (2) it uses oriented features, and (3) it accurately models most of the prominent aspects of primary visual cortex. The transform uses seven basic features (kernels), which may be regarded as three oriented edges, three oriented bars, and one non-oriented blob. Application of these kernels to non-overlapping seven-pixel neighborhoods yields six oriented, high-pass pyramid layers, and one low-pass (blob) layer.

Watson, Andrew B.↗

NASA NTRS1990

The method of constant stimuli is inefficient

Simpson (1988) has argued that the method of constant stimuli is as efficient as adaptive methods of threshold estimation and has supported this claim with simulations. It is shown that Simpson's simulations are not a reasonable model of the experimental process and that more plausible simulations confirm that adaptive methods are much more efficient that the method of constant stimuli.

Watson, Andrew B.↗

NASA NTRS1990

Gain, noise, and contrast sensitivity of linear visual neurons

Contrast sensitivity is a measure of the ability of an observer to detect contrast signals of particular spatial and temporal frequencies. A formal definition of contrast sensitivity that can be applied to individual linear visual neurons is derived. A neuron is modeled by a contrast transfer function and its modulus, contrast gain, and by a noise power spectrum. The distributions of neural responses to signal and blank presentations are derived, and from these, a definition of contrast sensitivity is obtained. This formal definition may be used to relate the sensitivities of various populations of neurons, and to relate the sensitivities of neurons to that of the behaving animal.

Watson, Andrew B.↗

NASA NTRS1990

Optimal displacement in apparent motion and quadrature models of motion sensing

A grating appears to move if it is displaced by some amount between two brief presentations, or between multiple successive presentations. A number of recent experiments have examined the influence of displacement size upon either the sensitivity to motion, or upon the induced motion aftereffect. Several recent motion models are based upon quadrature filters that respond in opposite quadrants in the spatiotemporal frequency plane. Predictions of the quadrature model are derived for both two-frame and multiframe displays. Quadrature models generally predict an optimal displacement of 1/4 cycle for two-frame displays, but in the multiframe case the prediction depends entirely on the frame rate.

Watson, Andrew B.↗

NASA NTRS1989

Supercomputer requirements for selected disciplines important to aerospace

Speed and memory requirements placed on supercomputers by five different disciplines important to aerospace are discussed and compared with the capabilities of various existing computers and those projected to be available before the end of this century. The disciplines chosen for consideration are turbulence physics, aerodynamics, aerothermodynamics, chemistry, and human vision modeling. Example results for problems illustrative of those currently being solved in each of the disciplines are presented and discussed. Limitations imposed on physical modeling and geometrical complexity by the need to obtain solutions in practical amounts of time are identified. Computational challenges for the future, for which either some or all of the current limitations are removed, are described. Meeting some of the challenges will require computer speeds in excess of exaflop/s (10 to the 18th flop/s) and memories in excess of petawords (10 to the 15th words).

Peterson, Victor L.↗

NASA NTRS1989

Ideal Resampling Of Discrete Sequences

Spectral information preserved to extent possible. Technique developed to shrink or expand discrete input sequence of numbers into output sequence in manner preserving input spectrum up to Nyquist limit of smaller of two sequences. In case of expansion, technique also prevents introduction of spurious frequencies not present in input. While applicable to many kinds of data sampled at regular interval, particularly useful in processing and enhancement of images, where discrete sequences spatially ordered sets of picture-element brightness values. Image coarsened by resampling to reduce number of picture elements while preserving as much as possible of original image information. Used to generate intermediate images at small intervals between frames, thereby suppressing appearance of jerky motion caused by sudden jumps between frames at low frame rate.

Watson, Andrew B.↗

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.↗