Showing papers on "Scale-invariant feature transform published in 1988"

Application of the generalised Hough transform to corner detection

Abstract: A new approach to corner detection is described which is based on the generalised Hough transform. The approach has the advantage that it can be used when objects have curved sides or blunt corners, as frequently happens with food products; in addition, it can be tuned for varying degrees of corner bluntness. The method is inherently sensitive: we have shown how it may be optimised for accuracy in the measurement of object dimensions and orientation.

On The Sensitivity Of The Hough Transform For Object Recognition

Abstract: Object recognition from sensory data involves, in part, determining the pose of a model with respect to a scene. A common method for finding an object's pose is the generalized Hough transform, which accumulates evidence for possible coordinate transformations in a parameter space whose axes are the quantized transformation parameters. Large clusters of similar transformations in that space are taken as evidence of a correct match. A theoretical analysis of the behavior of such methods is presented. The authors derive bounds on the set of transformations consistent with each pairing of data and model features, in the presence of noise and occlusion in the image. Bounds are provided on the likelihood of false peaks in the parameter space, as a function of noise, occlusion, and tessellation effects. It is argued that haphazardly applying such methods to complex recognition tasks is risky, as the probability of false positives can be very high. >

Shape detection using the adaptive Hough transform

Abstract: A computationally efficient implementation of the Hough transform, called the Adaptive Hough Transform is described. The method uses a coarse to fine search strategy to find peaks in the parameter space using a small accumulator array of fixed size. Its storage requirements are dramatically lower than those for the conventional Hough transform and it is also faster computationally when high dimensional parameter spaces are used. Shape detection using both natural representations of curves and hyperplane representations is discussed and the effectiveness of the method is demonstrated experimentally by applying it to the detection of parametric shapes in images of industrial objects.

Application of the Projected Hough Transform in Picture Processing

Abstract: P.V.C. Hough (1962) observed that in binary pictures lines correspond to maxima of the Radon transform. For this reason the Radon (or Hough) transform is frequently used in picture processing applications. Since the amount of data to be processed in performing this transformation can become very large, numerous approaches were proposed in the literature for data reduction by means of projection. In the present paper conditions are given for a generalized version of the Radon transform to yield useful results.

Yet Another Line Parametrization for Hough Transform

Abstract: The crucial problem of finding efficient line parametrizations when using the Hough transform to identify straight lines in edge enhanced images is discussed in this paper. A derivate of the parametrization by image edge intersections [Wa185] is investigated. It offers significantly reduced time and space requirements because of a relatively small and compact accumulator an algorithm to fill the accumulator which generates sharp peaks in parameter space facilitating cluster detection, which uses only integer arithmetic without multiplications and divisions, which can efficiently be implemented by parallel hardware, and an algorithm to combine the results produced in subimages into which the image is subdivided so that Hough transform and cluster detection in a proportionally smaller accumulator can be applied to these subimages separately and — where possible — in parallel.

On a Fast Piece-Wise Linear Hough Transform PLHT and its Application.

Abstract: Replacing the Hough calculation of the trigonomeric functions, sine and cos0,by the piece-wise linear Hough function(PLH), the basic cost for the sine , cose and the multiplications is removed. The PLH function is directly introduced from the usual Hough function. The PLH function inherits the basic properties of the usual Hough function from the view point to extract the line patterns from the pattern space. The computing cost of the PLH transform was reduced to about 1/6 of that of the usual Hough transform. It was also investigated that an additional property of the PLH transform contributes to reduce the memory cost to about 70 % of the usual Hough transform. Hough transform is one of the important methods to extract line patterns from the noisy and unclustered points of the image. As the edge or line patterns are the essential features in several industrial vision systems, it is practically required to make the Hough transform efficient from the view point of the computing and memory costs. A new fast Hough transform algorithm is introduced and its application is shortly presented in this paper. From this point of view, it is important to reduce the computing cost to utilize the Hough transform in several application. One of the factors in order to realize this cost reduction is to reduce the number of Hough calculations with respect to the number of edge points and and resolution numbers of the parameter space. It is still expectative to decrease the computing cost for the core Hough calculation defined by eq.(l). x.cos 0 + y.sin 8 (1) * =€I : perpendicular angle from x-axis JJ : the length of the perpendicular line In this paper, replacing Hough calculation of the trigonometric functions by the piecewise linear Hough(PLH) function which is composed of m pieces of line segments, the basic cost for the sine , cose and multiplications can be removed. It is shown in section 3 that the PLH function inherits the basic properties of the usual Hough function from the view point of to extract line patterns from the pattern space, and a few modifications of the pattern behavior in pattern space are also presented. In section 4, an new algorithm of piece-wise linear Hough transform(PLHT) is introduced, and some experimental results of PLHT are presented to demonstrate the reduction of the computing cost using an image of industrial engine parts. 2. Piece-wise Linear Hough Function 2.

Optimization of the Generalized Hough Transform

Abstract: Many image processing problem require curve detection. These include vision directed automation, remote control of vehicles, biomedical applications and so on. The Hough transform[1] [2] is a technique for detecting straight lines within a noisy image and later adapted for the detection of circles, ellipses and other analytically defined shapes. This method has been modified by D.H.Ballard[3] for detecting arbitrary shapes, which is called generalized Hough transform.

Recognition of Nonrigid Objects Using the Generalized Hough Transform

Abstract: Important visual cues in images are edges and line contours The application of feature detection algorithms does not necessarily provide a complete segmentation into closed contours A medium level step to combine characteristic local primitives to global structures is required The Hough Transform is well known as a robust transform which converts the global pattern detection problem into a local problem: finding evident maxima in parameter space, which is equivalent to detecting global structures in image space It is a model-based approach which assumes that the contours to be detected are a priori represented by an exact model curve However, objects in images are often slightly deformed, which impedes their successful recognition using the usual Hough techniques

Fast Hough Transform On A Mesh Connected Processor Array

Abstract: Hough transform is an effective method for the detection of the shape of object boundaries in image pattern analysis. Since the Hough transform is very computation intensive, it is essen-tial to parallelize the computation. However, an effective parallel algorithm is harder to obtain because it requires global informa-tion. In this paper we present an efficient parallel Hough transform algorithm for the detection of straight lines using mesh connected processor arrays. While other parallel algo-rithms take either 0(n2) or 0(n2) time, where n is the number of distinct values of a parameter and N is the number of edge pixels, our algorithm takes 0(n) time.