Sparse Representation Wavelet Based Classification

Ngo, Long H.; Luong, Marie; Sirakov, Nikolay; Le-Tien, Thuong; Guérif, Sébastien; Viennet, Emmanuel

doi:10.1109/icip.2018.8451538

Cited by 8 publications

(10 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, these methods are not resistant to training samples degradation by high-amplitude noise or occlusions. A recent approach, namely Sparse Representation Wavelet based Classification (SRWC) [20] was developed in the wavelet domain, taking advantage of the wavelet-promoted sparsity and leading to better discrimination. To improve the image classification accuracy, the SRWC exploits the sparse representation of the features described by the complementary information from the low-frequency sub-band of the wavelet coefficients.…”

Section: Related Workmentioning

confidence: 99%

“…As a consequence, the SRWC method exhibited higher accuracy compared to the contemporary SRbased methods in [9], [11], [12]. In [20], an overcomplete dictionary is constructed by mapping the training samples into the wavelet domain. To determine the membership of the query test sample, its sparse vector is first computed through an l 1 -norm minimization problem.…”

Section: Related Workmentioning

confidence: 99%

“…Inspired by the work in [20] and taking advantage of the SR of sparsity-promoting wavelets in the AQ, the present study develops a novel SRC method in the QW domain.…”

Section: Related Workmentioning

confidence: 99%

“…Usually, the aforementioned SRC method is applied to the spatial domain. A recent approach was proposed by Ngo et al [20] and performed in the sparsity-promoting wavelet domain to enhance the sparsity level of the codes and their discrimination ability. Despite its successful applications to classification [21], the discrete wavelet transform (DWT) is restricted by its lack of shift-invariance [22].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Image Classification Based on Sparse Representation in the Quaternion Wavelet Domain

et al. 2022

Self Cite

View full text Add to dashboard Cite

In this study, we propose a novel sparse representation learning method in the Quaternion Wavelet (QW) domain for multi-class image classification. The proposed method takes advantages from: i) the QW decomposition, which promotes sparsity and provides structural information about the image data while allowing approximate shift-invariance, to extract meaningful features from low-frequency QW subbands, ii) the dimensionality reduction method using Principal Component Analysis (PCA) for reducing the complexity of the problem, and iii) the sparse representation of the generated QW features to efficiently learn and capture the meaningful and compact information of this data. After the QW decomposition, the features extracted from low-frequency image sub-bands information are projected, by the PCA, into a new feature space with lower dimensionality. The features extracted from the training samples are used to construct a dictionary, while the features of the test samples are sparsely coded for the classification step. The sparse coding problem is formulated in a QW Least Absolute Shrinkage and Selection Operator (QWLasso) model applying quaternion l 1 minimization. A novel Quaternion Fast Iterative Shrinkage-Thresholding Algorithm (QFISTA) is developed to solve the QWLasso model. The experiments conducted on various public image datasets validated that the proposed method possesses higher accuracy, sparsity, and robustness in comparison with several contemporary methods in the field including Neural Networks.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

“…Inspired by the work in [20] and taking advantage of the SR of sparsity-promoting wavelets in the AQ, the present study develops a novel SRC method in the QW domain.…”

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Image Classification Based on Sparse Representation in the Quaternion Wavelet Domain

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…Thus, NNs are useful to modeling complex relationships between inputs and outputs or to find patterns in data with an indeterminant number of parameters or with parameters whose features cannot easily be modeled using traditional regression and statistical methods. In this kind of abilities, the NNs are analogous to the sparse representation classification methodology [7] which learns and uses large dictionaries.…”

Section: Introductionmentioning

confidence: 99%

Classification with Stochastic Learning Methods and Convolutional Neural Networks

2020

Self Cite

View full text Add to dashboard Cite

The present study illustrates a convolution neural network designed to classify skin lesion images to benign and malignant melanoma. The convolutional neural network consists of 4 convolution layers, rectified linear activation function and a softmax classifier. The new architecture is experimented to extract features with the following learning methods: Adaptive Moment Estimation and Stochastic Gradient Descent which optimize the weight changing functions for quicker convergence to minimal error between the ground truth and the real output. To further strengthen the accuracy and accelerate the convergence of classification, random noise is added to the neural network learning methods. To validate the theoretical concepts, experiments are conducted with 3800 skin lesion images from the ISIC2018 dataset. At the end of the paper, the results obtained by the convolutional neural network with noisy Stochastic Gradient Descent and noisy Adaptive Moment Estimation are compared with other contemporary classifiers including convolution neural network. Table 3 reports the accuracy of 95.23%, obtained by the new classifier. This value establishes the supremacy of the newly proposed neural network over the other contemporary neural network classifiers.

show abstract

Quaternion Wavelet Transform (QWT)

Hitzer

2023

Trends in Mathematics

View full text Add to dashboard Cite

Sparse Representation Wavelet Based Classification

Cited by 8 publications

References 16 publications

Image Classification Based on Sparse Representation in the Quaternion Wavelet Domain

Image Classification Based on Sparse Representation in the Quaternion Wavelet Domain

Classification with Stochastic Learning Methods and Convolutional Neural Networks

Quaternion Wavelet Transform (QWT)

Contact Info

Product

Resources

About