Showing papers in "Image and Vision Computing in 1992"

TL;DR: A new approach is proposed which works on range data directly and registers successive views with enough overlapping area to get an accurate transformation between views and is performed by minimizing a functional which does not require point-to-point matches.

TL;DR: The curvedness is a positive number that specifies the amount of curvature, whereas the shape index is a number in the range [−1, +1] and is scale invariant, which captures the intuitive notion of ‘local shape’ particularly well and can be mapped upon an intuitively natural colour scale.

TL;DR: The analysis develops measurement models from the fundamental principles of laser radar operation and identifies the nature and cause of key problems that affect measurements from this class of sensors.

TL;DR: A technique for constructing shape representation from images using free-form deformable surfaces, which results in a wide range of applications: reconstruction of smooth isolated objects such as human faces, reconstruction of structured objectssuch as polyhedra, and segmentation of complex scenes with mutually occluding objects.

TL;DR: The results indicate that existing methods of superquadric parameter estimation are quite sensitive to noise in range values and also do not work well for ‘rough’, coarse-surfaced objects, however, for low-noise real range images of objects with ‘smooth’ surfaces, numerical features derived from estimated super quadric parameters can be used with a binary tree classifier to infer qualitative shape properties with about 80% reliability.

TL;DR: The use of the Γ-convergence theory to approximate the functional to be minimized by elliptic functionals, which are more tractable and of relevance to vision applications are suggested.

TL;DR: A general analytical model is developed for the structural grey-value transitions in an image which is the superposition of elementary model functions and it will be shown that this parametric model agrees fairly well with real image intensities.

TL;DR: It is shown how to recover the position and orientation of a pair of known coplanar conies from a single perspective image and concludes that using such methods enables accurate pose determination for arbitrary plane curves.

TL;DR: A multi-scale representation of grey-level shape that makes explicit features in scale-space as well as the relations between features at different levels of scale is presented, which gives a qualitative description of the image structure that allows for extraction of significant image structure in a solely bottom-up data-driven manner.

TL;DR: An analogical representation of space and time is described that supports the formation of an event level description giving a local interpretation of what the objects are doing in the scene and how to form the global level conceptual descriptions.

TL;DR: A linear algorithm for recovering shape information of Lambertian surfaces from the shading information inherent in a single 2D image is presented and results for noisy synthetic images as well as real images are presented.

TL;DR: The edge preserving restoration of piecewise smooth images is formulated in terms of a probabilistic approach, and a MAP estimate algorithm is proposed which could be implemented on a hybrid neural network.

TL;DR: It is shown that a novel technique for computing integer prefix sums on meshes, augmented with a reconfigurable bus system (reconfigurable meshes), affords us fast algorithms to compute low-level descriptors including perimeter, area, histogram, and median row.

TL;DR: By combining enhanced fuzzy relaxation and association graph techniques, the mthod integrates global inter-object relations and local object attributes to obtain more reliable matching.

TL;DR: The use of the Radon transform is evaluated as a means whereby the transition from edge image data to symbolic representation may be realized and a methodology is suggested for object recognition and the recovery of the viewer object correspondence.

TL;DR: The curvatures estimated on a set of ventricle-shaped models are significantly closer to their analytical values than those assessed by four other methods.

TL;DR: It is proved that, in general, knowledge of the true size and shape does not yield a sufficient number of constraints to determine uniquely 3D orientation and position, and extra constraints must be acquired from other sources and fused with these constraints to obtain a unique solution to the 3D-location-estimation problem.

TL;DR: A robust algorithm for segmentation of noisy images using a neural network to represent a segmented image in such a way that the region type of each pixel simply corresponds to the neuron state is presented.

TL;DR: A new contour extraction method based on organization of the image data using perceptual grouping rules, and is therefore largely domain independent, allowing the filtering of noisy edges and line segments regardless of their strength.

TL;DR: A new method is presented which robustly deals with errors in edge detector output based on the concept of a chamfer map, which closes gaps while preserving edge structure and maintains surface connectedness correctly.

TL;DR: A new technique for finding boundaries and junctions, the half boundary finding algorithm, which uses a combination of morphological feature detection and a novel boundary following procedure to overcome the criticisms of previous edge finders.

TL;DR: An unconventional approach to range image understanding is proposed, unconventional in the sense that object matching and recognition is performed before the image is segmented, and the problem of finding the optimal point-wise mapping is formulated as one of constrained optimization.

TL;DR: The unsuitability of existing stereo matchers for this task is described in the light of recent work within UCL-PS, and a number of possible solutions are outlined.

TL;DR: The purpose of this note is to show that this new paradigm can be used for computing the Hough transform—the most widely used technique for detecting lines and curves in binary or grey level images.

TL;DR: Four algorithms that modify the histogram equalization algorithm, and extend its capability to a larger range of histograms, are presented and the effects of the algorithms on peaks or Gaussions having various means, standard deviations and population sizes are discussed.

TL;DR: A theoretical framework is demonstrated within which it is possible to construct descriptors for curves which do not vary with viewpoint, known as invariants, which make it possible to recognise plane curves, without explicitly determining the relationship between the curve reference frame and the camera coordinate system.

TL;DR: Progress in providing the Oxford Autonomous Guided Vehicle (AGV) with range sensing and model-based recognition of classes of industrial artifacts is described and a system for pose recognition of generic models from range-imagery is described.

TL;DR: A comparative evaluation of some fast parallel thinning algorithms for elongated shapes on an overlapped shared memory multiprocessor architecture (OSMMA) is presented and the algorithms implemented include the own thinning technique based on the concept of object edges width compression that achieves efficient results while significantly decreasing execution timings.