Design of FIR bilevel Laplacian-of-Gaussian filter
Journal
Signal Processing
Journal Volume
82
Journal Issue
4
Pages
677-691
Date Issued
2002
Author(s)
Horng, Ji-Hwei
Abstract
The Laplacian-of-Gaussian (LoG) filter is an optimal edge detector, but it is computationally inefficient. In this paper, we propose the bilevel Laplacian-of-Gaussian (BLoG) filter to approximate the LoG filter. This approximation is formulated as an optimization problem and solved by the gradient descent algorithm. Only two multiplications per pixel are required in convolving with the BLoG filter. The computation is fast and it produces satisfactory results. The applications of 1D BLoG filter to corner detection and 2D BLoG filter to edge detection are also presented. © 2002 Elsevier Science B.V. All rights reserved.
Subjects
Corner detection; Edge detection; Gradient descent; Laplacian-of-Gaussian; McClellan transformation
Other Subjects
Algorithms; Edge detection; Gradient methods; Mathematical transformations; Optimization; Problem solving; McClellan transformation; FIR filters
Type
journal article
File(s)![Thumbnail Image]()
Loading...
Name
64.pdf
Size
512 KB
Format
Adobe PDF
Checksum
(MD5):18dcb62f9057ef2c3d6e4cfc807ec2a5