Edge and Curve Detection: Further Experiments

Rosenfeld, Azriel; Thurston, Mark; Lee, Yung-Han

doi:10.1109/t-c.1972.223573

Cited by 102 publications

(14 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several methods for detecting edges based on gray-level gradient and angular changes in digital curves were proposed in [193,195,196,197]. Other solutions for line-drawing images include methods for detecting corners in a chain-coded plane curve [73,74].…”

Section: High Curvature Pointsmentioning

confidence: 99%

Local Invariant Feature Detectors: A Survey

Tuytelaars

Mikolajczyk

2007

FNT in Computer Graphics and Vision

1,146

573

View full text Add to dashboard Cite

In this survey, we give an overview of invariant interest point detectors, how they evolved over time, how they work, and what their respective strengths and weaknesses are. We begin with defining the properties of the ideal local feature detector. This is followed by an overview of the literature over the past four decades organized in different categories of feature extraction methods. We then provide a more detailed analysis of a selection of methods which had a particularly significant impact on the research field. We conclude with a summary and promising future research directions.

show abstract

Section: High Curvature Pointsmentioning

confidence: 99%

Local Invariant Feature Detectors: A Survey

Tuytelaars

Mikolajczyk

2007

FNT in Computer Graphics and Vision

1,146

573

View full text Add to dashboard Cite

show abstract

“…Working before the advent of the wavelet framework, Rosenfeld and coworkers suggested forming multiscale point-wise products [26][27][28]. This is intended to enhance multiscale peaks due to edges, while suppressing noise, Figure 1.…”

Section: Multiscale Productsmentioning

confidence: 99%

“…We also note the earlier work of Witkin [33], who developed ideas for multiscale characterization of signals. Working on image processing before the advent of wavelets, Rosenfeld and coworkers suggested an interesting idea, to form multiscale point-wise products [26][27][28]. This approach is intended to enhance multiscale peaks due to edges, while suppressing noise, by exploiting the multiscale correlation due to the presence of the desired signal in a direct (albeit nonlinear) way.…”

Section: Introductionmentioning

confidence: 99%

Analysis of Wavelet Transform Multiscale Products for Step Detection and Estimation

Sadler¹,

Swami²

1998

View full text Add to dashboard Cite

We consider discrete wavelet transform (DWT) multiscale products for detection and estimation of steps. Here the DWT is an overcomplete approximation to smoothed gradient estimation, with smoothing varied over dyadic scale, as developed by Mallat and Zhong. We show that the multiscale product approach, as first proposed by Rosenfeld for edge detection, is a nonlinear whitening transformation. We characterize the resulting non-Gaussian heavy-tailed densities. The results may be applied to edge detection with a false alarm constraint. The response to impulses, steps, and pulses is also characterized. A general closed-form expression for the Cramer-Rao bound (CRB) for discrete and continuous-time step-change location estimation in independent identically distributed non-Gaussian noise is developed, generalizing previous results. We consider location estimation using multiscale products, and compare results to the appropriate CRB.ii

show abstract

“…Another way of defining the precise boundaries of the blob is to project its surrounding edges downward through the pyramid, one level at a time, and at each level to locally adjust them so as to maximize the gray level gradient magnitude along the edges. However, if this simple method of edge adjustment is used, noise at the lower levels of the pyramid will cause the edges to break up, since they are not constrained to becontinuous (Rosenfeld, Thurston, & Lee, 1972). Better results are obtained if one starts with a connected region (the union of the pixels surrounded by the edges), projects it downward through the pyramid, and at each step, adjusts its border by adjoining or deleting pixels in such a way as to maximize the contrast around the border.…”

Section: Law Of Closure: Detecting Blobs and Ribbonsmentioning

confidence: 99%

Pyramid algorithms for perceptual organization

Rosenfeld

1986

Behavior Research Methods, Instruments, & Computers

View full text Add to dashboard Cite

Multiresolution (or pyramid) approaches to computer vision provide the capability of rapidly detecting and extracting global structures (features, regions, patterns, etc.) from an image. The human visual system also is able to spontaneously (or preattentively) perceive various types of global structure in visual input; this process is sometimes called perceptual organization. This paper describes a set of pyramid-based algorithms that can detect and extract these types of structure; included are algorithms for inferring three-dimensional information from images and for processing time sequences of images. If implemented in parallel on cellular pyramid hardware, these algorithms require processing times on the order of the logarithm of the image diameter.During the past few years, there has been increasing interest in the use of multiresolution (pyramid) image representations in image analysis and computer vision. Several research groups have designed or built imageprocessing machines basedon this approach. A pyramidstructured array of multiprocessors can perform many types of operations on an image (inputto the base of the pyramid, one pixel per processor), in time proportional to the log of the imagediameter. (For a recent collection of papers on pyramid methods in image processing and analysis, see Rosenfeld, 1984.) One of the most important potential applications of pyramidmachines is the fast detection and extraction of globalstructures (e.g., features, regions, patterns) in an image, by rapidlycombining information collected from many parts of the image. A number of pyramid-based methods of extracting global imagestructureshave been developed at the Center for Automation Research, and others have been proposed, as described below.The rapid detection of global structure in images is an important, but not well-understood, capability of the human visual system. Humans tend to spontaneously (or preattentively) perceive various typesof globalstructures in their visual input;this process is sometimes calledperceptual organization. About 50 years ago, the Gestalt psychologists formulated a set of principles, or laws, that describehow imageparts tend to group intoglobal structures (Wertheimer, 1958). These include the laws of similarity, proximity, good continuation, and closure. In this paper, I will showhow each of these typesof groupings can be detected using fast pyramid algorithms.Fast detection of globalstructure in an image seems to be an essential component of real-time perception. HuRequests for reprints should be mailed to:

show abstract

Edge and Curve Detection: Further Experiments

Cited by 102 publications

References 1 publication

Local Invariant Feature Detectors: A Survey

Local Invariant Feature Detectors: A Survey

Analysis of Wavelet Transform Multiscale Products for Step Detection and Estimation

Pyramid algorithms for perceptual organization

Contact Info

Product

Resources

About