Showing papers in "Pattern Recognition Letters in 1987"

TL;DR: A more general definition of features such as edges, shadows and bars is developed, based on an analysis of the phase of the harmonic components, showing that these features always occur at points of maximum phase congruency.

TL;DR: It is shown that a fuzzy group can be associated with a fuzzy graph in a natural way and some properties of fuzzy graphs are considered and the notions of eccentricity and center are introduced.

TL;DR: A new method for extracting depth information from 2-dimensional images is propsed using measurements of the degree of blur in an image which is only in parts sharply focused.

TL;DR: Algorithms used for extracting facial features such as head outline, location of eyes, eyebrows and mouth from a grey-scale image of the face are described, and knowledge of appropriate positions of features at a given resolution is used to guide searches at finer resolutions.

TL;DR: A high resolution image is found such that when simulating the imaging process the authors get low resolution images closest to the observed images, and simulated annealing is also used for the optimization process.

TL;DR: Two-dimensional Gabor filters are used to segment images into regions of specific spatial frequency or orientation characteristic and the images are transformed into a modulated narrowband signal whose envelope coincides with the region(s) whose characteristics the filter is tured to.

TL;DR: A new class of edge-preserving noise-cleaning filters are introduced which use both spatial and nearest-neighbor constraints on image pixels to smooth an image.

TL;DR: The main idea is to consider a fuzzy subset of an images as the nested family of its level-cuts, interpret this family as a body of evidence in the sense of Shafer, and any intrinsic parameter can be calculated as a mathematical expectation based on a probability density function.

TL;DR: Using Brodatz's textures, the proposed features are evaluated and compared with those suggested by Conners et al. (1984) to present a new approach to texture feature extraction from a cooccurrence matrix.

TL;DR: The operation of the MLP is described as a pattern recognition device in terms of a feature-space representation, which allows an understanding of how structure in the training data is represented internally in the machine.

TL;DR: A quantitative evaluation of Hough transform methods for detecting faint linear features in synthetic aperture radar images and techniques to locate points in the transform space and locate the end points of the corresponding lines are presented.

TL;DR: An effective but simply implemented strategy is devised which achieves speedup factors of up to ∼25 with real images and is highly robust against shape defects and partial occlusions, and in addition robustness is easily monitored in practical situations.

TL;DR: Psychophysical evidence suggests a primal sketch in human vision, rather like that proposed by Marr, although different in many details, that the human version does not have an address-based geometrical representation of space, and that the process of spatial representation may be by-passed for certain dense contour arrangements.

TL;DR: A method is presented for nonlinear inversion of operators that enables nonlinear restoration of image restoration techniques and it is empirically shown that the Conjugate Gradients algorithm converges quickly in this vector space and enables inverted of operators on big pictures with relatively little computation.

TL;DR: This work defines generic data structures for geometric shapes, along with robust rules for parsing the image geometry and performing a shape-motivated resegmentation, and successfully applies the system to the task of locating and outlining complex rectilinear cultural objects in aerial imagery.

TL;DR: Lateral histograms are found to be particularly valuable for finding small round objects and holes as well as for corner location, and methods for resolving ambiguities are described.

TL;DR: A linear and a spatial representation scheme are derived from binary curve relations that describe a curve by a simple chain code, called Ruli and the procedures reducing the resolution of both representations possess the length reduction property.

TL;DR: It is shown that in the presence of a simple geometrical constraint city block distance between n -dimensional lattice point is characterized by a condition used by Rosenfeld in an investigation of continuous functions on digital images.

TL;DR: A method for automatically detecting arterial bifurcations in successive angiogram frames and a proposed automated analysis system utilising this are described.

TL;DR: The relationship between the Markov mesh and nonsymmetric half-plane Markov chains leads to the conclusion that the NSHP Markov chain is more appropriate for applications.

TL;DR: It is found that the Hough transform is not a simple matched filter and that it has sub-optimal signal detection capability; however, sensitivity is improved by gradient weighting of points in parameter space, when it becomes proportional to image contrast.

TL;DR: The early results of research into developing algorithms to permit the counting and tracking of moving vehicles in real world scenes using the CLIP4 parallel image processor are outlined.

TL;DR: Euclidean distance transformations (EDT) and skeletons are used to give robust characterizations of shape and to represent multi-component objects by close approximations to the continuous medial axis.

TL;DR: An algorithm for automatic image segmentation using a ‘homogeneity’ measure and a “contrast” measure defined on the co-occurrence matrix of the image are described and the comparison of its performance with the existing ones are demonstrated.

TL;DR: This note presents a modified method of curve partitioning based on selecting points at which the distance of the curve from its chord(s) is a local maximum that seems to yield generally reasonable results.

TL;DR: It is found that noise affects estimation of edge orientation in a complex way, but this is simplified for those ‘circular’ operators which act in a strictly vectorial manner, so the distribution of edge orientations is closely gaussian for gaussian image noise.

TL;DR: It is argued that the most effective algorithms for automatic speech recognition derive their effectiveness from an ability to delay low-level decisions until higher level decisions (such as segmential identities and boundaries) have been made.

TL;DR: This representation of the Wigner Distribution for discrete images is shown to be specially adequate for processing of textured information and of spatially variant degraded images.

TL;DR: A methodology is described for reliably locating pre-defined features in constrained images using a geometric model incorporating knowledge about the images involved, and a grey-level model capable of identifying specific features in a grey level profile.

TL;DR: A theorem is presented which shows that statistics of cross ratios of groups of four collinear points are a projective invariant and the distance between normalized vectors is used as the similarity measure for texture discrimination.