This paper considers both the Hueckel operator and a recently introduced class of edge detectors implemented as two-dimensional infinite impulse response (IIR) or recursive digital filters. The latter were originally ...
详细信息
This paper considers both the Hueckel operator and a recently introduced class of edge detectors implemented as two-dimensional infinite impulse response (IIR) or recursive digital filters. The latter were originally developed on the basis of least mean-square Wiener spatial filtering concepts for an assumed stochastic model of edge structure in typical imagery data. These two schemes are compared with respect to performance, computational complexity and suitability for specific applications. Problems in making these comparisons are discussed. The implementations of both operators used for this comparison are described.
A new method is proposed for correcting possible errors in phonemic-strings which are obtained by some preprocessing of spoken words and then applying statistical decision to resulting features. The proposed method us...
详细信息
A new method is proposed for correcting possible errors in phonemic-strings which are obtained by some preprocessing of spoken words and then applying statistical decision to resulting features. The proposed method uses new distance measure and the available linguistic information in the form of rules of the following two types: rules obtained from the theory of word formation (from phonemes) in English, and rules needed to correct some error made by statical decision rules used earlier. In an experiment, involving 400 utterances and 40 words, an error correction rate of over 99 percent was achieved using the proposed error correction method.
Significant energy savings in the paper industry can result from characterization of paper pulps by direct measurement of fiber morphology with automatic methods. The various fiber properties are discussed for which a...
详细信息
Significant energy savings in the paper industry can result from characterization of paper pulps by direct measurement of fiber morphology with automatic methods. The various fiber properties are discussed for which automatic image analysis applied to microscope images of pulps can yield savings of materials and energy. computational algorithms applied to actual fiber data are shown to yield useful measurements for fiber length, curl, and potentially other properties. New fiber morphology measurement algorithms can include curl and length as special cases. Potentials for collaboration between the paper industry and the image analysis instrument industry are explored.
A key element in patternrecognition is the description of shape. The description of shape is facilitated by segmenting the boundary line at so-called critical points - corners (discontinuities in curvature), points o...
详细信息
A key element in patternrecognition is the description of shape. The description of shape is facilitated by segmenting the boundary line at so-called critical points - corners (discontinuities in curvature), points of inflection, and curvature maxima. Additional critical points are intersections and points of tangency. Algorithms are described for extracting such critical points in the presence of noise. An illustration is given showing how the critical points may be used in the development of a shape description system.
Straight-line edge segments are often used as key features in building scene models. This paper presents techniques for extracting edge segments from a digital picture. Results of line extraction on a few outdoor scen...
详细信息
Straight-line edge segments are often used as key features in building scene models. This paper presents techniques for extracting edge segments from a digital picture. Results of line extraction on a few outdoor scenes are also presented.
Sixty-three papers were presented at the conference held at Rensselaer Polytechnic Institute in June, 1977. Areas of coverage were pattern and image analysis, image models, medical applications, speech pattern recogni...
详细信息
Sixty-three papers were presented at the conference held at Rensselaer Polytechnic Institute in June, 1977. Areas of coverage were pattern and image analysis, image models, medical applications, speech patternrecognition, syntactic patternrecognition, industrial applications, shape analysis, scene analysis, multi-imageprocessing, statistical patternrecognition, and character and signature recognition.
The processes of associated storage and recall can be represented mathematically in the framework of linear mappings, termed associative mappings. Three kinds are here considered with the help of pictorial demonstrati...
详细信息
The processes of associated storage and recall can be represented mathematically in the framework of linear mappings, termed associative mappings. Three kinds are here considered with the help of pictorial demonstrations, which reveal the inherent pattern-processing properties of these mappings. The autoassociative mapping is shown optim reconstruct images from incomplete or noisy versions, when the correct image has been previously stored. The Novelty Filter has the ability to extract the ″novelty″ in a test pattern, with respect to stored patterns, as is revealed by a demonstration using pictures of different facial expressions. The third mapping, the heteroassociative mapping, is used in two demonstrations to classify patterns that have not been previously stored, showing its ability to interpolate between stored items.
A general adaptive patternrecognition and learning software system, suitable for a broad range of applications, has been developed. This system integrates state-of-the-art methods in patternrecognition and learning ...
详细信息
A general adaptive patternrecognition and learning software system, suitable for a broad range of applications, has been developed. This system integrates state-of-the-art methods in patternrecognition and learning in a software structure uniquely suited to the special characteristics of ballistic missile defense classication problems. The capability of this software was compared experimentally with that of an existing nonadaptive patternrecognition program. The experiments indicate that the adaptive software has advantages over nonadaptive software in adjusting to actual data characteristics, recognizing new classes of data, and learning rapidly, even with imperfect feedback.
The TICAS software package is designed for research in quantitative cytology. The user is guided by the programs to choose options and to sequence program modules such as recording of digitized cell images through the...
详细信息
The TICAS software package is designed for research in quantitative cytology. The user is guided by the programs to choose options and to sequence program modules such as recording of digitized cell images through the microscope, feature extraction, statistical feature evaluation, feature selection, and self-learning algorithms for automated cell recognition and classification. Examples for some typical analytical procedures are given.
Cluster analysis for patterns represented for sentences is investigated. The similarity between patterns is expressed in terms of the distance between their corresponding sentences. A weighted distance between two str...
详细信息
Cluster analysis for patterns represented for sentences is investigated. The similarity between patterns is expressed in terms of the distance between their corresponding sentences. A weighted distance between two strings is defined and its probabilistic interpretation given. The class membership of an input pattern (sentence) is determined according to the nearest neighbor of K-nearest neighbor rule. A clustering procedure on a sentence-to-sentence basis is proposed. A set of English characters is used to illustrate the proposed metric and clustering procedure.
暂无评论