Wavelet transforms on two-dimensional images

Schmeelk, John

doi:10.1016/s0895-7177(02)00238-8

Cited by 23 publications

(11 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Figure 3 illustrates a twodimensional discrete wavelet transform with three levels, each of which yields four sub-bands. The first level generates LL1 (low-low), HL1 (high-low), LH1 (low-high), and HH1 (high-high) [30]. The second level, LL1, is decomposed to generate LL2, LH2, HL2, and HH2, and the third level generates LL3, LH3, HL3, and HH3, which provide details of horizontal, vertical, and diagonal directions at the three decomposition levels.…”

Section: Discrete Wavelet Transformmentioning

confidence: 99%

Competent Radiogenomics Features for the Prediction of glioma Mutations

Tarek¹,

Rady²,

Badr³

2022

Preprint

View full text Add to dashboard Cite

Glioma is a type of brain tumor, mostly diagnosed through biopsy or surgery. The prediction of glioma is still challenging to offer the required accurate solution for the diagnosis process. The challenge lies more at the edge of finding a solid feature representation rather than employing well-tuned classifiers. This paper proposes a solution to improve glioma prediction and diagnosis by proposing a rich combination of radiomic and genomic features. A novel feature extraction mechanism is proposed relying on the triple combination CNN-GLCM-DWT for MRI images and PCA for the genomics data. Several classical and deep learning classifiers are trained on the data. The proposed feature extraction is tested for validation on a dataset of 422 patients with low- or high-grade gliomas to predict three mutation statuses. Experiments prove the competency of the proposed mechanism when trained with an SVM classifier, which achieves high accuracy values of 97.5%, 96%, and 92% for the three mutations IDH1 mutation status, 1p/19q codeletion, and MGMT promotor methylation status, respectively. CNN as a deep learning method comes second with accuracy values of 95%,94.2%, and 88%, while the Random Forest comes last with values of87.5%, 80.7%, and 76%. Compared to existing related works that apply a single feature extraction method, the proposed triple combination CNN-GLCM-DWT feature extraction excels in classification accuracyfor the three mutation types by approximately 3.7%, 3.3%, and 10% inaccuracy. It is faster in average classification time by approximately 50%.

show abstract

Section: Discrete Wavelet Transformmentioning

confidence: 99%

Competent Radiogenomics Features for the Prediction of glioma Mutations

Tarek¹,

Rady²,

Badr³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…The division of the LL sub-band is further carried out into the four previous stated bands. The detailed component of the image is represented by LH, HL, and HH [63]. The compaction of the image is incremented with the incrementation of levels, but the quality of the desired approximate image is decreasing.…”

Section: Discrete Wavelet Transformmentioning

confidence: 99%

An Efficient Methodology for Brain MRI Classification Based on DWT and Convolutional Neural Network

Fayaz

Torokeldiev²,

Turdumamatov

et al. 2021

Sensors

View full text Add to dashboard Cite

In this paper, a model based on discrete wavelet transform and convolutional neural network for brain MR image classification has been proposed. The proposed model is comprised of three main stages, namely preprocessing, feature extraction, and classification. In the preprocessing, the median filter has been applied to remove salt-and-pepper noise from the brain MRI images. In the discrete wavelet transform, discrete Harr wavelet transform has been used. In the proposed model, 3-level Harr wavelet decomposition has been applied on the images to remove low-level detail and reduce the size of the images. Next, the convolutional neural network has been used for classifying the brain MR images into normal and abnormal. The convolutional neural network is also a prevalent classification method and has been widely used in different areas. In this study, the convolutional neural network has been used for brain MRI classification. The proposed methodology has been applied to the standard dataset, and for performance evaluation, we have used different performance evaluation measures. The results indicate that the proposed method provides good results with 99% accuracy. The proposed method results are then presented for comparison with some state-of-the-art algorithms where simply the proposed method outperforms the counterpart algorithms. The proposed model has been developed to be used for practical applications.

show abstract

“…Here, two-dimensional discrete wavelet transform [14] was employed to remove the noise signal in the 2D fingerprints. By experimental analyses, bior3.7 basis wavelet function and 1-level decomposition were selected.…”

Section: Preprocessmentioning

confidence: 99%