Showing papers on "Feature (computer vision) published in 1983"

Model-Based Three-Dimensional Interpretations of Two-Dimensional Images

Abstract: ACRONYM is a comprehensive domain independent model-based system for vision and manipulation related tasks. Many of its submodules and representations have been described elsewhere. Here the derivation and use of invariants for image feature prediction is described. Predictions of image features and their relations are made from three-dimensional geometric models. Instructions are generated which teli the interpretation algorithms how to make use of image feature measurements to derive three-dimensional size, structural, and spatial constraints on the original three-dimensional models. Some preliminary examples of ACRONYM's interpretations of aerial images are shown.

Processing translational motion sequences

Abstract: A procedure for processing real world image sequences produced by relative translational motion between a sensor and environmental objects is presented. In this procedure, the determination of the direction of sensor translation is effectively combined with the determination of the displacement of image features and environmental depth. It requires no restrictions on the direction of motion, nor the location and shape of environmental objects. It has been applied successfully to real-world image sequences from several different task domains. The processing consists of two basic steps: feature extraction and search. The feature extraction process picks out small image areas which may correspond to distinguishing parts of environmental objects. The direction of translational motion is then found by a search which determines the image displacement paths along which a measure of feature mismatch is minimized for a set of features. The correct direction of translation will minimize this error measure and also determine the corresponding image displacement paths for which the extracted features match well.

"Compensatory articulation" under conditions of reduced afferent information: a dynamic formulation.

Abstract: A well-established feature of speech production is that talkers, faced with either anticipated or unanticipated perturbations. can spontaneously adjust the movement patterns of articulators such th...

Interactive imagery and affective judgments improve face-name learning in the elderly.

Abstract: Groups of elderly adults were taught to learn name-to-face associations using one of three different techniques. In a control group (no image) participants were taught for each face-name pair to select a prominent facial feature and to transform the surname into a concrete word. Persons in a second group (image) additionally were taught to employ interactive imagery to form an association between the prominent feature and the transformed name. The third group (image + judgment) was treated the same as the second except that these individuals were also taught to judge the pleasantness of the image association that was formed. As predicted, improvement following instruction was minimal when no image association was formed but strong when interactive imagery was used. Moreover, those persons in the image + judgment group remembered more names than those in the image group and showed less forgetting on a measure of delayed recall. In addition to replicating and extending the findings of previous research with a different sample, the present study demonstrates that semantic orienting tasks can be used to enhance the retention of visual image associations as well as the simpler stimuli used in prior research.

The characterization of the arrangement of feature centroids in planes and volumes

Abstract: The quantitative characterization of the arrangement of microstructural features is important for understanding the properties or function of materials or tissues. In this paper, the well‐known analysis of nearest neighbour distances is modified in order to separate clusters from a background of randomly disposed features and to characterize significant parameters of such sets. In addition, simple procedures are indicated to detect and to characterize regular arrangements, and computer simulated examples are used to demonstrate the efficiency of the suggested algorithms. Each step of the procedures can be used for three‐dimensional arrangements as well.

Feature Article-Statistical Analysis of Simulation Output Data

Abstract: We present a state-of-the-art survey of statistical analyses for simulation output data of a single simulated system. The various statistical problems associated with output data analyses such as start-up bias and determination of estimator accuracy are described in detail. We then discuss the best available techniques for addressing these problems, as well as topics for future research. The paper concludes with a discussion of how developments in simulation languages, computer graphics, and computer execution speed may affect the future of output analyses.

Computing the optic flow: The MATCH algorithm and prediction☆

Abstract: Optic flow is the vector field induced on the retina by the movement of features in the external world. We address the stimulus-matching problem of approximating the optic flow with a velocity vector field inferred from hypothesized matching of features in successive “retinal snapshots.” We here present an algorithm, MATCH, for solving this problem based on the assumptions of smooth motion, (almost) rigid bodies, and nonsparse distribution of feature points on surfaces. We explore the performance of the algorithm first with simulated data and then with actual photographs, demonstrating the utility of prediction based on sequences of more than two frames.

3DPO: A Three-Dimensional Part Orientation System.

Abstract: A system that recognizes objects in a jumble, verifies them, and then determines some essential configurational information, such as which ones are on top, is presented. The approach is to use three-dimensional models of the objects to find them in range data. The matching strategy starts with a distinctive edge feature, such as the edge at the end of a cylindrical part, and then “grows” a match by adding compatible features, one at a time. (The order of features to be considered is predetermined by an interactive, off-line feature-selection process.) Once a sufficient number of compatible features has been detected to allow a hypothesis to be formed, the verification procedure evaluates it by comparing the measured range data with data predicted according to the hypothesis. When all the objects in the scene have been hypothesized and verified in this manner, a configuration-understanding procedure determines which objects are on top of others by analyzing the patterns of range data predicted from all the hypotheses. Experimental results of the system’s performance in recognizing and locating castings in a bin are presented.

Monocular depth perception from optical flow by space time signal processing

Abstract: A theory of monocular depth determination is presented. The effect of finite temporal resolution is incorporated by generalizing the MarrHildreth edge detected operator - V2G(r) where V2 is the Laplacian and G(r) is a two-dimensional Gaussian. The constraint that the edge detection operator in space-time should produce zero-crossings at the same place in different channels, i.e. at different resolutions of the Gaussian, led to the conclusion that the Marr-Hildreth operator should be replaced by - LI2G(r, t) where E2 is the d'Alembertian V2- (1/u2)(a2/at2) and G(r, t) is a Gaussian in space-time. To ensure that the locations of the zerocrossings are independent of the channel width, G(r, t) has to be isotropic in the sense that the velocity u appearing in the defintion of the d'Alembertian must also be used to relate the scales of length and time in G. However, the new operatior - LI2G(r, t) produces two types of zero-crossing for each isolated edge feature in the image I(r, t). One of these, termed the 'static edge', corresponds to the position of the image edge at time t as defined by V2I(r, t) = 0; the other, called a 'depth zero', depends only on the relative motion of the observer and object and is usually found only in the periphery of the field of view. When an edge feature is itself in the periphery of the visual field and these zeros coincide, there is an additional cross-over effect. It is shown how these zero-crossings may be used to infer the depth of an object when the observer and object are in relative motion. If an edge feature is near the centre of the image (i.e. near the focus of expansion), the spatial and temporal slopes of the zeros crossing at the static edge may be used to infer the depth, but, if the edge feature is in the periphery of the image, the cross-over effect enables the depth to be obtained immediately. While the former utilizes sharp spatial and temporal resolution to give detailed three-dimensional information, the cross-over effect relies on longer integration times to give a direct measure of the time-to-contact. We propose that both mechanisms could be used to extract depth information in computer vision systems and speculate on how our theory could be used to model depth perception in early visual processing in humans where there is evidence of both monocular perception of the environment in depth and of looming detection in the periphery of the field of view. In addition it is shown how a number of previous models are included in our theory, in particular the directional sensor proposed by Marr & Ullman and a method of depth determination proposed by Prazdny.

Determining velocity maps by spatio-temporal neighborhoods from image sequences

Abstract: Velocity information is not only important for determining velocities and trajectories of objects but also important as a cue for image segmentation. This paper presents a method of computing velocity information from a sequence of images. It uses velocity estimates of the prominent feature points as reliable initial estimates and propagates them efficiently to other points using a constraint relation between neighboring points. The result obtained from the first two frames is used for efficient analysis of subsequent frames. The presented method allows objects to translate and rotate in the 3-D world and may occlude one another. It is assumed, however, that objects are rigid and their motion is smooth. Experimental results arc shown for several synthetic and real world images.

Post-Decisional Reevaluation of Choice Alternatives

Abstract: In a free-choice experiment, subjects chose between two pairs of alternatives, each containing a positively and a negatively valued feature. The results confirmed dissonance theory in that the negative feature of the chosen alternatives increased in value from before to after the choice. However, the positive feature of the rejected alternative decreased only slightly. Implications of the findings for the theory are discussed.

Feature navigation system and method

Abstract: A method of navigating an airborne vehicle using observation of linear features such as roads, railroads, rivers or the like. Maps of linear features in several discrete areas along an intended flight path are prepared and entered into an on-board computer memory. The vehicle typically uses an inertial navigation system during flight from the origin to the first mapped area. Imaging sensors and on-board processing detect linear features as the vehicle approaches or passes over them. The sensed feature pattern network is compared to the reference mapped feature network to update the inertial navigation system to generate guidance commands. Frequent updates at other mapped areas keep the vehicle on course. This system is also capable of navigation updates from single linear features when the flight path is fairly well known, and of tracking a linear feature for terminal guidance. In these cases, only a small number of feature parameters need to be stored, rather than an area map. Linear feature navigation can be used by any manned or unmanned airborne vehicle which has the on-board capability to detect the required features.

System for detecting orientation of articles

Abstract: Data relating to the distances between each of all the feature points of the article to be checked and the other feature points thereof is stored in a memory in advance. The image of the article is picked up by a camera, and the resulting image data is stored in another memory. With use of a computer, data relating to the distances of at least two of the feature points of the article from other feature points thereof is obtained from the image data. The two feature points are identified by checking the data relating to the distances of the two feature points from other feature points with reference to the stored data relating to the distances between each feature point and the other feature points. The orientation of the article is determined based on the direction through the two identified feature points.

Perceptual organization as a basis for visual recognition

Abstract: Evidence is presented showing that bottom-up grouping of image features is usually prerequisite to the recognition and interpretation of images. We describe three functions of these groupings: 1) segmentation, 2) three-dimensional interpretation, and 3) stable descriptions for accessing object models. Several principles are hypothesized for determining which image relations should he formed: relations are significant to the extent that they are unlikely to have arisen by accident from the surrounding distribution of features, relations can only be formed where there are few alternatives within the same proximity, and relations must be based on properties which are invariant over a range of imaging conditions. Using these principles we develop an algorithm for curve segmentation which detects significant structure at multiple resolutions, including the linking of segments on the basis of curvilinearity. The algorithm is able to detect structures which no single-resolution algorithm could detect. Its performance is demonstrated on synthetic and natural image data.

Pattern features extracting apparatus and method

Abstract: A plurality of different mask pattern data which are indicated by values obtained from Hermitte's polynomials of different degrees weighted by a Gaussian function are stored in mask memories, respectively. Each mask pattern data stored in each of the mask memories is convolved by a sum-of-product circuit together with unknown input pattern data. A plurality of extracted pattern feature data are stored in pattern feature memories.

Virtual addressing for E-beam lithography

Abstract: A technique performed in a fixed address particle beam lithographic system where the writing is performed in the normal manner for writing a pattern, for example, a stripe on a resist having a selected feature width except that an additional row of alternate pixels is written either before or after the selected feature is written. The alternate pixels, when the resist is developed, will provide a feature width of approximately 1/2 a pixel wider than the selected feature width due to blurring of the latent image caused by scattering of the particle beam within the resist. Thus, the resolution of selectable feature widths is enhanced with little or no loss of throughput. The same technique can also be utilized to lengthen a feature by 1/2 a pixel width. The technique is disclosed primarily in a raster scan machine but also disclosed is the technique in a vector scan machine. Also disclosed is a flow chart showing the invention used while preparing the data to be written by the machine.

VLSI architectures for feature extraction and pattern classification

Abstract: In statistical methods for image processing and pattern classification, large-scale matrix computations are often performed over huge image data bases. A class of matrix arithmetic networks is proposed for implementing the Foley-Sammon feature extraction algorithm and for generating linear discriminant vectors in pattern classification. Such VLSI feature extractors and pattern classifiers are in high demand in real-time artificial intelligence applications. Performances of the proposed VLSI image analyzers are compared with conventional software approaches using a uniprocessor computer.

Model-based interpretation of range imagery

Abstract: This paper describes a model-based approach to interpreting laser range imagery. It discusses the object modeling, model-driven prediction, and feature-to-model matching aspects of the problem. The model objects are represented by a viewpoint-independent volumetric model based on generalized cylinders. Predictions of 3-D image features and their relations are generated from object models on multiple levels. These predictions give guidance for goal-directed shape extraction from low level image features. Interpretation proceeds by comparing the extracted image features with object models in a coarse to fine hierarchy. Since the 3-D information is available from the range image, the actual measurements are used for feature-to-model matching. A limited prototype system has been developed, preliminary results on prediction and interpretation are shown, and future research directions are discussed.

Configurations of a chemically ordered continuous random network to describe the structure of GeSe2 glass

Abstract: It is proposed that the structure of GeSe2 glass is characterized by a chemically ordered continuous random network. Within this context it is suggested that 4-atom planar rings exist and can explain an unusual feature in the Raman spectrum. Neutron scattering data are presented which supportthis model.

Method and apparatus for automatic optical focusing on an optically discernible feature on an object

Abstract: A method for the automatic fine focusing of an image scanning densitometer, micrometer, or television camera or the like, after an initial manual focus has been made on the edge of a discernible feature on the surface of an object of interest. A plurality of at least three, and preferably five, scans are made across the image of the feature at a corresponding plurality of focus positions, and the variations in light intensities are noted at each of a plurality of locations in each scan so that an edge value may be determined for each focus position. If plotted, edge values as a function of focus position appears as a second order curve of the form, edge value=ae2 +be+c, from which the values of the coefficients "a" and "b" may be determined by the least squares best fit method. When the derivative of the curve equation is made equal to zero to correspond to a zero slope or bottom of the curve, the best focus position is readily determined to that the precise focus may be adjusted to that position.

Expansion of the left atrial appendage is a distinctive cross-sectional echocardiographic feature of congenital defect of the pericardium.

Abstract: A case report of a congenital defect of the pericardium in which the cross-sectional appearance of a prominent left atrial appendage is similar to the previously reported angiographic diagnostic appearance.

Feature IC'S for Digivision TV Sets

Abstract: As interface between the processors two types of digital bus systems are used: A video bus system giving access to the composite video signal and to the processed luminance and color difference signals and a control bus system which is used to control the different processors.The clock rate of the video bus system is fourtimes color subcarrier frequency and the clock rate of the control bus is approx. 170 kHz. All new systems are designed to interface directly with the two bus system of the digital TV chip set.

Self mutilation as a feature of Addison's disease.

Abstract: Psychiatric symptoms reported in Addison's disease tend to be non-specific: lethargy and lassitude are the most common. Addison himself described a patient who developed depression eight years and again six months before his death from the disease.' We report a case of longstanding self mutilitation that regressed completely after treatment for previously unsuspected adrenocortical insufficiency.

A doubly layered, genetic penetrance learning system

Abstract: The author's original state-space learning system (based on a probabilistic performance measure clustered in feature space) was effective in optimizing parameterized linear evaluation functions. However, more accurate probability estimates would allow stabilization in cases of strong feature interactions. To attain this accuracy and stability, a second level of learning is added, a genetic (parallel) algorithm which supervises multiple activations of the original system. This scheme is aided by the probability clusters themselves. These structures are intermediate between the detailed performance statistics and the more general heuristic, and they estimate an absolute quantity independently of one another. Consequently the system allows both credit localization at this mediating level of knowledge and feature interaction at the derived heuristic level. Early experimental results have been encouraging. As predicted by the analysis, stability is very good.

Analysis of data from a single epidemic1

Abstract: Summary A model with nonrandom latent and infectious periods is suggested for epidemics in a large community. This permits a relatively complete statistical analysis of data from the spread of a single epidemic. An attractive feature of such models is the possibility of exploring how the rate of spread of the disease depends on the number of susceptibles and infectives. An application to smallpox data is included.