China
Africa
Explore chapters and articles related to this topic
A Study on Feature Extraction and Classification Techniques for Melanoma Detection
Published in P. Madhumathy, M. Vinoth Kumar, R. Umamaheswari, Machine Learning and IoT for Intelligent Systems and Smart Applications, 2021
S. Poovizhi, T. R. Ganesh Babu, R. Praveena, J. Kirubakaran
Shearlet transform uses anisotropic dilation and orientation to precisely capture the geometric edges. The continuous shearlet transform is represented as ψa,s,t(x)=a−34ψ(A−1B−1(x−t))SHψf(a,s,b)=<f,ψa,s,t>,a>0,s
Image Processing Concepts
Published in Manas Kamal Bhuyan, Computer Vision and Image Processing, 2019
Shearlet transform can be used to detect singularities, step edges, and corner points of the images. As shown in Figure 2.38, shearlet has more directional bands in the frequency domain as compared to curvelet for the same set of decomposition levels. Shearlet transform is a combination of Laplacian pyramid Laplacian pyramid and a set of shearing filters. First, the input source image is fed to the Laplacian pyramid which gives one low-frequency band, and remaining bands belong to high-frequency contents. The low-frequency band can be further fed into the Laplacian pyramid. The number of times the low-frequency band is decomposed using the Laplacian pyramid is called the “decomposition level of the system.” For the reconstruction of the input image, the sum of all shearing filters responses and low-frequency bands are fed into the inverse Laplacian pyramid. Figure 2.39 shows a 2-level decomposition scheme of multi-scale and multi-directional NSST. Figure 2.40 shows the different frequency features obtained using the shearlet transform.
Chapter 10: Wavelets
Published in Abul Hasan Siddiqi, Mohamed Al-Lawati, Messaoud Boulbrachene, Modern Engineering Mathematics, 2017
Abul Hasan Siddiqi, Mohamed Al-Lawati, Messaoud Boulbrachene
The shearlet and curvelet [38] approach is designed to deal with directional and anisotropic features typically present in natural images, and have the ability to accurately and efficiently capture the geometric information of edges.
An efficient Peaceman–Rachford splitting method for constrained TGV-shearlet-based MRI reconstruction
Published in Inverse Problems in Science and Engineering, 2019
Tingting Wu, Wenxing Zhang, David Z. W. Wang, Yuehong Sun
In numerical experiments, the linear operator in (1.1) consists of p rows of the n-by-n () matrix representing the full 2D DFT. The p selected rows specify the selected frequencies at which the measurements in f are collected. The sampling ratio is defined by p / n. Accordingly, the scanning duration will be short if the sampling ratio p / n is small. In magnetic resonance imaging, there are some freedoms to select the rows corresponding to a certain frequencies. Herein, we use the pseudo radial sampling patterns (see e.g. Figures 2 and 3) which can simulate randomness in acquisition for demonstration purpose. For the shearlet transform, we exploit the three-scale case with 29 subbands. Concretely, one subband for low frequency and 28 subbands for high frequency. The package for implementing shearlet transform are available at http://www.mathematik.uni-kl.de/imagepro/software/.
QR code-based non-linear image encryption using Shearlet transform and spiral phase transform
Published in Journal of Modern Optics, 2018
Ravi Kumar, Basanta Bhaduri, Bryan Hennelly
Recently, a new approach called the discrete Shearlet transform (DST) has been proposed to improve the previous methods (36). DST is a new discrete representation for multi-scale directionality, which is based on two types of ability: the ability to use the power of multi-scale methods and the ability to capture the geometry of multidimensional data (35). It is an affine system that contains a single mother Shearlet function, which is parameterized by scaling, shear and translation parameters with the shear parameter capturing the direction of the singularities (36). Mathematically, the Shearlet transform is implemented using a Laplacian pyramid scheme and directional filtering. For an image I, the Shearlet transform is a mapping as (35):
Medical image fusion based on multi-scale decomposition using hybrid deep learning network model
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2023
Syed Munawwar, P. V. Gopi Krishna Rao
Dogra and Kumar (2022) developed a novel hybrid technique for image fusion. The shearlet transform is used for image decomposition on multi-modal images. In multidirectional orientations, the original image’s texture information is captured by this shearlet transform. Then, it decomposes these paired images. The base layer is combined with these weights to create unified base layers. Utilizing the inverse shearlet transform yields the final fusion result.