Showing papers in "Pattern Recognition in 1990"

TL;DR: The simulation results indicate that the proposed segmentation algorithm yields the most accurate segmented image on the color coordinate proposed by Ohta et al.

TL;DR: There still is a great gap between human reading and machine reading capabilities, and a great amount of further effort is required to narrow-down this gap, if not bridge it.

TL;DR: Promising results are obtained when applying the Texture Spectrum to classify four of Brodatz's natural images.

TL;DR: Estimation theory is used to derive a new approach to the clustering problem, a unification of centroid and mode estimation, achieved by considering the effect of spatial scale on the estimator, which is a multiresolution method which spans a range of spatial scales.

TL;DR: A state-of-the-art of methodology and algorithms of fuzzy sets in the field of pattern recognition and clustering techniques are discussed and a problem of cluster validity expressed in terms of clustering indices is addressed.

TL;DR: Recognition experiments with a prototype system for a variety of complex printed documents shows that the proposed system is capable of reading different types of printed documents at an accuracy rate of 94.8–97.2%.

TL;DR: This study will describe a structural model of fingerprints, based on local structural relations among features, and an associated automated recognition system which addresses the limitations of existing fingerprint models.

TL;DR: Results obtained show that direct feature statistics such as the Bhattacharyya distance are not appropriate evaluation criteria if texture features are used for image segmentation, and that the Haralick, Laws and Unser methods gave best overall results.

TL;DR: The line density in thespace is homogenized, the efficiency of utilization of the space is increased, and a stable normalization is obtained for partially irregular shape variations.

TL;DR: Taking advantage of the orthogonality property, a systematic feature selection method for choosing an appropriate number of the Zernike features is developed, based on computing a measure of the information content differences of features of different classes.

TL;DR: A new spatial knowledge representation 2D C-string with accompanied cutting mechanism and a set of spatial operators are proposed that is more characteristic of spatial knowledge in images and is more efficient in representation and manipulation of images.

TL;DR: Une liste de 308 termes relatifs a la vision par ordinateur avec leur definition avecLeur definition est proposee.

TL;DR: It is demonstrated how it is possible to achieve a globally consistent MAP estimate by iterative label replacement without sacrificing the representational capacity of the label process.

TL;DR: This paper evaluates a number of corner detectors and addresses some questions, such as number of responses to a single corner and noise sensitivity, about the performance of the algorithms.

TL;DR: The techniques covered are segmentation of text and contextual recognition, both of which are required when considering text recognition of documents, and the various techniques for storing dictionary information are compared.

TL;DR: A complementary approach which is to determine the stationary background first is proposed, so that once the background is known, the moving objects can be easily located.

TL;DR: A human face profile recognition system with learning capability is developed and one key point of the imaging system is the use of back-lighting so that the outline curve can be extracted easily.

TL;DR: A graph-theoretic optimization method is used to recognize partially occluded objects from a 2-D image through the use of maximal cliques and a weight matching algorithm to improve the computational speed.

TL;DR: Simulated annealing is used to find a near-optimal partitioning with respect to each of several clustering criteria for a variety of simulated data sets, which can then be used to determine the “best” clustering criterion for the multi-sensor fusion problem with a given fusion operator.

TL;DR: An experimental workstation, a vector-based recognition method, and an intelligent interaction function which are proposed in order to enhance input speed are presented.

TL;DR: Size and species grading of fish (e.g. on board a fishing boat) might in future be done entirely automatically using image analysis by creating a data bank of shapes from photographs of seven different types of fish on a computer.

TL;DR: A new dynamic strip algorithm for fitting a curve with lines that is insensitive to a scale and orientation changes and obtained by merging the results from several runs of the algorithm with different minimum strip widths is presented.

TL;DR: It is shown that incremental search can be implemented as a sequence of invocations of a previously published non-incremental algorithm, and a new incremental search algorithm is presented which finds the next nearest neighbor more efficiently by eliminating redundant computations.

TL;DR: A survey of the state of the art in on-line handprinted character recognition technology, with particular attention to the effects of handwriting variability and constraints imposed, and to the human factors involved.

TL;DR: A totally automatic non-parametric clustering algorithm and its application to unsupervised image segmentation and systematic methods for automatic selection of an appropriate histogram cell size are developed and discussed.

TL;DR: Experiments on some Chinese characters show that the proposed new techniques are reliable and time saving compared with the direction method.

TL;DR: A general adaptive model unifying existing models for pattern learning, unifying many existing models (including that of neural nets), is proposed, and suggests how various propositional object (class) descriptions might be generated based on the outputs of the learning processes.

TL;DR: The development of algorithms for the machine recognition of optically captured Arabic characters and their isolation from the printed text and an algorithm for separation of individual characters are discussed.

TL;DR: Local metrics are described and are generalized to the multi-class case and Properties of various possible generalizations are compared, and the results of a simulation study are described.

TL;DR: This paper demonstrates that higher classification accuracies can be obtained from the same training set by using a combination of decision trees and by reaching a consensus using Dempster and Shafer's theory of evidence.