We propose a biologically motivated method, called nonclassical receptive field (non-CRF) inhibition (more generally, surround inhibition or suppression), to improve contour detection in machine vision. Non-CRF inhibition is exhibited by 80% of the orientation-selective neurons in the primary visual cortex of monkeys and has been shown to influence human visual perception as well. Essentially, ...

A computational framework for extracting (1) edges with an arbitrary proole function and (2) keypoints such as corners, vertices and terminations is presented. Using oriented lters with even and odd symmetry we combine their convolution outputs to oriented energy resulting in a uniied representation of edges, lines and combinations thereof. We derive an "edge quality" measure which allows to te...

Much of our knowledge about motion perception has been obtained by studying bars moving within apertures. When viewed within an ambiguously oriented aperture such as a circle, bars appear to move orthogonal to their orientation. We demonstrate that if the local orientation of the aperture edges is altered, a direction consistent with the edge orientation is seen. Indeed, the perceived direction...

We introduce a structured image compression scheme based on a u = v + w model, where the orignal image u is decomposed between a sketch v and a residue w. The sketch contains the main edge curves while the information carried by the residue is mostly composed by microtextures. This split of the meaningful information allows the use of a representation adapted to each structure. The edge curve g...

Edge detection and edge enhancement techniques play an essential role in interpreting potential field data. This paper describes the application of various edge detection techniques to gravity data in order to delineate the edges of subsurface structures. The edge detection methods comprise analytic signal, total horizontal derivative (THDR), theta angle, tilt angle, hyperbolic of tilt angle (H...

Implementation of the Convex Edge Segment Grouper Qingmin Shi Department of Computer Science and Engineering University of South Florida 4202 E. Fowler Avenue, Tampa, FL 33620 Email: [email protected] Abstract| This report describes the implementation of convex line segment grouper proposed by David W. Jacobs[9]. The inputs of this algorithm are line segments obtained by edge detector and strai...

Edge tracking by surface detection in the spatiotemporal continuum or nding correspondence by not loosing it Abstract We present a framework and algorithms for edge tracking in image sequences. Edges are tracked as discontinuity surfaces in the 3-D spatiotemporal continuum spanned by the frames of the sequence. An eecient data structure for coding the topological and metrical structure of 3-D p...

Edge loading causes clinical problems for hard-on-hard hip replacements, and edge loading wear scars are present on the majority of retrieved components. We asked the question: are the lines of action of hip joint muscles such that edge loading can occur in a well-designed, well-positioned acetabular cup? A musculoskeletal model, based on cadaveric lower limb geometry, was used to calculate for...

In this paper we present a brief overview of the processing in the primary visual cortex, the multi-scale line/edge and keypoint representations, and a model of brightness perception. This model, which is being extended from 1D to 2D, is based on a symbolic line and edge interpretation: lines are represented by scaled Gaussians and edges by scaled, Gaussian-windowed error functions. We show tha...

In cutting thin sections of dense tissue such as bone, much depends on the integrity of the glass knife edge. The method of breaking the glass by first scoring it into strips has certain limitations. These strips have irregular edges be-T]~xT-Fm. 1. Plan of piece of plate glass. cause they break along the strain lines introduced by the glass cutter and it follows that the cutting edge of the kn...