Showing papers in "Pattern Recognition in 1993"

TL;DR: Attempts have been made to cover both fuzzy and non-fuzzy techniques including color image segmentation and neural network based approaches, which addresses the issue of quantitative evaluation of segmentation results.

TL;DR: The invariants from second- and third-order moments are derived and shown to be complete and to be invariant under general affine transformation and used for recognition of affine-deformed objects.

TL;DR: This method provides an unbiased estimate of a binarized version of the image in an information theoretic sense by minimizing the cross entropy between the image and its segmented version.

TL;DR: The invention is a system failure monitoring method and apparatus which learns the symptom-fault mapping directly from training data and takes advantage of temporal context and estimate class probabilities conditioned on recent past history.

TL;DR: The results show that facial features can be located very quickly with Adequate or better fit in over 80% of the images with the proposed system.

TL;DR: A simple mathematical model is developed which computes fingerprint local ridge orientation from core and delta positions, which provides an intelligent tool for resolving ambiguities due to the periodic nature of orientation in algorithms for interpreting fingerprint patterns.

TL;DR: The theoretical analysis of the influence of noise on the location and the orientation of an edge is presented and reveals that the accuracy of the proposed approach is virtually unaffected by the additive noise.

TL;DR: Improved moment invariants computed using the shape boundary only are proposed, which tremendously reduces computations and are mathematically proved to be invariant to scaling, translation, and rotation.

TL;DR: The reject option proposed by Chow is extended by defining an ambiguity reject option and a distance reject option, which can be defined in a parametric as well as in a non-parametric way.

TL;DR: An important conclusion about the present method is that the Foley-Sammon optimal set of discriminant vectors is a special case of the set of optimal discriminant projection vectors.

TL;DR: Algorithms are identified which are best suited for an automatic fingerprint recognition system operating on low quality images, and suggestions are made concerning the most suitable algorithms in each of the processing steps.

TL;DR: It is shown how such spatio-chromatic features can be extracted using multi-scaled filtering and correlation methods which capture the variations of colour over space in ways which encode important image features not extracted by techniques which separate colour, texture and shape into separate channels.

TL;DR: As an area with sufficient challenge, considerable application potential, and extensive test data, handwritten Chinese character recognition is a good testbed for new recognition algorithms and enables connectionist models to be compared with more traditional recognition methods.

TL;DR: A new method for the recognition of arbitrary two-dimensional shapes based on string edit distance computation is described, which is invariant under translation, rotation, scaling and partial occlusion.

TL;DR: A new approach to digital implementation of continuous-scale mathematical morphology based on discretization of evolution equations associated with continuous multiscale morphological operations is presented.

TL;DR: This algorithm dynamically searches for the globally optimal solution on a minimum solution tree which is a subtree of the solution tree used in the traditional branch and bound algorithm.

TL;DR: A camera system is designed that can obtain stereo information in a single shot and through a single lens that is equivalent to two images obtained by two aligned cameras with exactly the same optical properties.

TL;DR: The proposed method for the purification of characteristics uses both foreground and background characteristics; that is, ridge and valley minutiae, which allows a fast and effective purification for ridge breaks.

TL;DR: This paper presents and discusses a method which attempts, in the off-line case, to recover part of the lost script dynamics, to reconstruct one of the most likely trajectories followed by the writer while drawing characters.

TL;DR: In a study based on 15 distinct Brodatz textures it is found that the tuning process although computationally intensive converges efficiently; the mean classifier values of the classifier for a particular texture at different orientation and different scales are tightly clustered.

TL;DR: Curvature properties have been extracted after thinning the smoothed character images and filtering the thinned images using a Gaussian kernel and the unknown samples are classified using a two-stage feed forward neural net based recognition scheme.

TL;DR: A method is introduced to combine and jointly optimize recognition and image normalization in optical character recognition algorithms based on pseudo two-dimensional (2D) hidden Markov models (HMMs) that provides a maximum likelihood estimate of the transformation parameters that can be used by higher level modules in an intelligent document recognition system as an aid in the recognition process.

TL;DR: An overview is presented of the principal on-line techniques for handwriting preprocessing and word normalization, covering the major difficulties encountered and the various approaches usually used to resolve these problems.

TL;DR: A technique to determine optimal polygons of digital curves is proposed that determines the longest possible line segments with the minimum possible error using the L 1 norm.

TL;DR: The approach is to obtain the congruent transformations which make the transformed image a good match with the original image, and the directional correlation of edge features is used to evaluate how well the original and transformed images match.

TL;DR: The suggested hybrid continuous-discrete approach is well suited to applications in which continuous surfaces are given in digitized form, computation time must be relatively short and an approximate solution is acceptable, and can be used to efficiently obtain a good initial approximation as an input to other algorithms.

TL;DR: A heuristic solution is proposed to lessen the inherent ambiguity of the principal axes method for object orientation and show excellent percentage classification success rates by using only a small number of normalized moments as elements of the feature vector.

TL;DR: A defined ETPL( k ) graph grammar is a subfamily of a well-known edNLC graph grammar being a strong formalism for pattern representation and generates a wider class of languages than the ones presented in the previous papers.

TL;DR: A Kohonen self-organizing neural network is used as a feature-vector quantizer and a “best match only” character-based recognizer performs better than a ‘best matchonly’ stroke-based Recognizer at the cost of a substantial increase in computation.

TL;DR: In this article, a multi-layer constraint satisfaction neural network (CSNN) is proposed for image recognition, which is capable of identifying the components of a brain in a preliminarily segmented image.