A Local Scale Selection Scheme for Multiscale Area Integral Invariants
Author(s)
Wang, Bin
Gao, Yongsheng
Sun, Changming
Blumenstein, Michael
La Salle, John
Griffith University Author(s)
Year published
2016
Metadata
Show full item recordAbstract
Area integral invariant (AII) is a functional obtained by performing integral operations on the closed planar contour of a shape via the convolution with disc kernels. This shape descriptor is insensitive to noise and robust with respect to occlusions. AII intrinsically introduces the notion of scale using the size of kernel radius. However how to select an optimal scale remains unresolved. In this paper, we propose a local scale selection scheme for generating multiscale area integral invariants. For the same scale level, the disc kernel size is not fixed and varies with the contour point where the disc is centered. This ...
View more >Area integral invariant (AII) is a functional obtained by performing integral operations on the closed planar contour of a shape via the convolution with disc kernels. This shape descriptor is insensitive to noise and robust with respect to occlusions. AII intrinsically introduces the notion of scale using the size of kernel radius. However how to select an optimal scale remains unresolved. In this paper, we propose a local scale selection scheme for generating multiscale area integral invariants. For the same scale level, the disc kernel size is not fixed and varies with the contour point where the disc is centered. This scheme also provides a scale assignment for emphasising the features extraction at finer scales. The strong discriminative power of the multiscale area integral invariant derived from the proposed scale selection scheme has been validated through experiments on very challenging leaf image retrievals.
View less >
View more >Area integral invariant (AII) is a functional obtained by performing integral operations on the closed planar contour of a shape via the convolution with disc kernels. This shape descriptor is insensitive to noise and robust with respect to occlusions. AII intrinsically introduces the notion of scale using the size of kernel radius. However how to select an optimal scale remains unresolved. In this paper, we propose a local scale selection scheme for generating multiscale area integral invariants. For the same scale level, the disc kernel size is not fixed and varies with the contour point where the disc is centered. This scheme also provides a scale assignment for emphasising the features extraction at finer scales. The strong discriminative power of the multiscale area integral invariant derived from the proposed scale selection scheme has been validated through experiments on very challenging leaf image retrievals.
View less >
Conference Title
2016 INTERNATIONAL CONFERENCE ON DIGITAL IMAGE COMPUTING: TECHNIQUES AND APPLICATIONS (DICTA)
Subject
Pattern recognition
Image processing