This conference proceedings contains 93 papers. All papers are indexed separately. Topics covered include: patternrecognition theory;industrial applications;edge processing;structural patternrecognition and clusteri...
详细信息
This conference proceedings contains 93 papers. All papers are indexed separately. Topics covered include: patternrecognition theory;industrial applications;edge processing;structural patternrecognition and clustering;applied imagery patternrecognition;curve and shape description;scene analysis;medical applications;texture and mathematical foundations;two dimensional and three dimensional shape description and matching;scene segmentation and interpretation;image coding and curve approximation;and, remote sensing and time-varying imagery.
An optimal linear discriminant function algorithm is introduced which minimizes the error rate of internal samples. It is applied to two sets of data. One data set is imaginary and the other is actual data drawn from ...
详细信息
An optimal linear discriminant function algorithm is introduced which minimizes the error rate of internal samples. It is applied to two sets of data. One data set is imaginary and the other is actual data drawn from the medical field.
A system for detecting and classifying images on astronomical plates is described. In particular, the very faint and barely resolved images of 24th magnitude galaxies are shown to be detectable and distinguishable fro...
详细信息
A system for detecting and classifying images on astronomical plates is described. In particular, the very faint and barely resolved images of 24th magnitude galaxies are shown to be detectable and distinguishable from stellar images of similar brightness. The algorithms and other implementation considerations used in the construction of the Faint Object Classification and Analysis System, FOCAS, are given. Detection and classification at 24th magnitude is achieved with N//2 presensitized IIIaJ plates exposed for one hour in the KPNO 4-m telescope prime focus camera. For a digital 6000 multiplied by 6000 pixel image made from the central square area 37 ft on a side approximately 16,000 objects are found to 24. 7 mag on such plates. Since the resulting catalog of objects is 3 or 4 magnitudes deeper than existing surveys a variety of internal tests must be used to assess the accuracy and completeness of the list of objects. The tests and simulations made to support these claims are also discussed.
Future generation x-ray computed tomography scanners will be characterized by their ability to record simultaneously a sufficient number of x-ray projections to allow reconstructions of multiple adjacent cross section...
详细信息
Future generation x-ray computed tomography scanners will be characterized by their ability to record simultaneously a sufficient number of x-ray projections to allow reconstructions of multiple adjacent cross sections of the object under study. An ability to repeat the entire data collection procedure with great rapidity, allowing many scan passes per second, should encourage research and diagnostic studies of moving organs such as the heart and lungs in truly three dimensions and in real time. A combined series of algorithmic, software, special-purpose computer architecture, and hardware implementation studies have demonstrated significant progress toward computed tomography reconstruction processing rates of 10**9 to 10**1**0 arithmetic operations per second.
A general method is presented for generation of stochastic texture fields from a priori given second order statistics. A counter example to Julesz conjecture is shown. This conjecture (1962) stated that human eye cann...
详细信息
A general method is presented for generation of stochastic texture fields from a priori given second order statistics. A counter example to Julesz conjecture is shown. This conjecture (1962) stated that human eye cannot discriminate two textures having the same second order statistics.
Beginning studies are described aimed at defining a set of features to be used with the Spatial Gray Level Dependence Method (SGLDM) which will measure image characteristics believed to be used by humans in the discri...
详细信息
Beginning studies are described aimed at defining a set of features to be used with the Spatial Gray Level Dependence Method (SGLDM) which will measure image characteristics believed to be used by humans in the discrimination of textures. In particular a procedure is described which will allow the SGLDM to be used to measure the unit cell size of textures and the coarseness of textures.
A description is provided of a particular computed Radiography system as it evolved from a computed Tomography system. A comparison of image performance is made between this system and standard film screen radiography...
详细信息
A description is provided of a particular computed Radiography system as it evolved from a computed Tomography system. A comparison of image performance is made between this system and standard film screen radiography. Some of the imageprocessing techniques used to construct and enhance the radiographic images are described.
It is proposed that the curves commonly encountered in picture processing can be described by a sequence of arcs linked at critical points. A statistical model is developed which assumes that the locations of the crit...
详细信息
It is proposed that the curves commonly encountered in picture processing can be described by a sequence of arcs linked at critical points. A statistical model is developed which assumes that the locations of the critical points are statistically independent. The model is used to develop and analyze algorithms for determining such critical points.
A segmentation is done on a monochrome 256 multiplied by 256 image using only texture information generated by local extreme measurements. The number of extrema of various sizes are counted over a local region and the...
详细信息
A segmentation is done on a monochrome 256 multiplied by 256 image using only texture information generated by local extreme measurements. The number of extrema of various sizes are counted over a local region and these numbers are used to separate the regions. The use of texture edge information in the local averaging process is discussed.
A recursive algorithm is given to directly compute the fast Fourier Transform (FFT) of a two-dimensional image array. The algorithm builds the FFT of a 2**m by 2**m matrix left bracket f right bracket by first mapping...
详细信息
A recursive algorithm is given to directly compute the fast Fourier Transform (FFT) of a two-dimensional image array. The algorithm builds the FFT of a 2**m by 2**m matrix left bracket f right bracket by first mapping its entries to form a new matrix left bracket f//1 right bracket , in a manner analogous to that used in one-dimensional FFT's. The FT of left bracket f right bracket is then found by computing the FT of all 2 by 2 submatrices of left bracket f//1 right bracket and recursively combining them into higher level submatrices until the 2**m by 2**m level is reached. This algorithm is shown to be computationally faster than the conventional method based on one-dimensional FFT's. The main features of this method, however, are its simplicity of updating the FFT of an image array when some of the array elements have been modified, and its suitability for implementation on a multiprocessor computer system. Applications of this algorithm to imageprocessing are discussed and illustrative examples are given.
暂无评论