Non-subsampled Shearlet Domain Research Articles

Multi-scale cross spectral coherence and phase spectral distribution based measurement in non-subsampled shearlet domain for classification of brain tumors

Computerized assessment of brain tumor identification and discrimination process from magnetic resonance imaging is of immense concern for improved investigation, growth rate prediction, treatment planning and monitoring of clinical trials. Due to extensive diversification of tumors and complex outline interactions between sub-regions, it is of a great challenge to extract significant features. The proposed approach starts with Grey Wolf Optimization based Gabor Filtering for enhancement process and spatial fuzzy c-means clustering algorithm for segmentation purpose. Following this, we apply non-sub-sampled shearlet transformation to investigate multi-scale space-frequency representation of tumor data. In order to measure the hierarchical correlation pattern of tumors, interlayer cross spectral coherence function is analyzed in terms of some significant indices such as fundamental global mean, secondary mean pair, span of secondary mean pair, percentage of samples having values greater than upper secondary mean pair and percentage of samples having values smaller than lower secondary mean pair. Additionally, topological profile of tumor getsassessed from its phase spectrum distribution pattern. Regarding this, some phase spectrum distribution-based quantifiers such as phase-sign change counter, phase change frequency, figure of merit and standard deviation of phase sign change location are introduced as the significant indices. These feature quantifiers have been fed to a deep neural network model and hybrid AdaBoost classifier to discriminate the state of benignancy/malignancy of tumors. Experimental results demonstrate that the overall accuracy attained by the proposed model are 99.67 %, 98.92 %, 96.89 %, 97.82 %, 96.98 %, 99.54 %for the datasets of Harvard University, BraTS −2018, −2019, −2020, −2021 and Kaggle respectively. Apart from this, promising sensitivity rate of the model helps us to avoid the risk of false negativity. Altogether, these qualities make the proposed model superior to the state of art techniques.

Glioma classification via MR images radiomics analysis

Accurate glioma classification before surgery is of the utmost important in clinical decision making and prognosis prediction. In this paper, we investigate the impact of multi-modal MR image fusion for the differentiation of low-grade gliomas (LGG) versus high-grade gliomas (HGG) via integrative analyses of radiomic features and machine learning approaches. A set of 80 histologically confirmed gliomas patients (40 HGG and 40 LGG) obtained from the MICCAI BraTS 2019 data were involved in this study. To achieve this work, we propose to combine T1 with T2 or FLAIR modality in the non-subsampled shearlet domain. Firstly, the pre-processed source MR images are decomposed into low-frequency (LF) and several high-frequency (HF) sub-images. LF sub-images are fused using the proposed weight local features fusion rule while HF sub-images are combined based on the novel sum-modified-laplacian technique. Experimental results demonstrate that the proposed fusion approach outperformed the recent state-of-the-art approaches in terms of entropy and feature mutual information. Subsequently, a key radiomics signature was retrieved by the least absolute shrinkage and selection operator regression algorithm. Five machine learning classifiers were established and evaluated with the retrieved dataset, then with the fused dataset using tenfold cross-validation scheme. As a result, the random forest had the highest accuracy of 96.5% with 21 features selected from the raw data and 96.1% with 16 features selected from the fused data. Finally, the experimental findings confirm that the proposed aided diagnosis framework represents a promising tool to aid radiologists in differentiating HGG and LGG.

NSST domain CT–MR neurological image fusion using optimised biologically inspired neural network

Diagnostic medical imaging plays an imperative role in clinical assessment and treatment of medical abnormalities. The fusion of multimodal medical images merges complementary information present in the multi-source images and provides a better interpretation with improved diagnostic accuracy. This paper presents a CT–MR neurological image fusion method using an optimised biologically inspired neural network in nonsubsampled shearlet (NSST) domain. NSST decomposed coefficients are utilised to activate the optimised neural model using particle swarm optimisation method and to generate the firing maps. Low and high-frequency NSST subbands get fused using max-rule based on firing maps. In the optimisation process, a fitness function is evaluated based on spatial frequency and edge index of the resultant fused image. To analyse the fusion performance, extensive experiments are conducted on the different CT–MR neurological image dataset. Objective performance is evaluated based on different metrics to highlight the clarity, contrast, correlation, visual quality, complementary information, salient information, and edge information present in the fused images. Experimental results show that the proposed method is able to provide better-fused images and outperforms other existing methods in both visual and quantitative assessments.

CT image denoising using multivariate model and its method noise thresholding in non-subsampled shearlet domain

In today era, computed tomography (CT) is one of the exceptionally proficient crucial devices in medical science for the clinical reason. The consistent improvement and broad utilization of computed tomography in medical science has uplifted the harmfulness of higher dose to the patient. Low radiation dose may prompt expanded noise and artifacts, which can influence the radiologists’ judgment. Therefore, we propose a method based on new shrinkage function in the nonsubsampled shearlet domain (NSST). In the proposed algorithm, method noise on multivariate shrinkage model is utilized viably by using stein's unbiased risk estimate and linear expansion of thresholds (SURE-LET) concept. To verify the execution of the proposed method, the qualitative and quantitative evaluations are performed. The results are evaluated over the both real noisy CT image and by adding Gaussian noise in real CT image and as well as on low complexity zoomed objects of noisy CT images. The results are also tested by some standard execution measurements, for example, PSNR, SSIM, ED, and DIV. The experimental results confirmed that proposed method is giving improved results in most cases.

Edge-preserving image denoising using a deep convolutional neural network

Abstract This paper introduces a novel denoising approach making use of a deep convolutional neural network to preserve image edges. The network is trained by using the edge map obtained from the well-known Canny algorithm and aims at determining if a noisy patch in non-subsampled shearlet domain corresponds to the location of an edge. In the first step of the proposed denoising algorithm, we use the non-subsampled shearlet transform to decompose the noisy image into a low-frequency subband and a series of high-frequency subbands. Subsequently, 3D blocks are formed by stacking 2D blocks of high-frequency subbands along a specific direction. Each 3D patch is then fed to the trained deep convolutional neural network to determine if it belongs to the edge-related class or not. Finally, the NSST (non-subsampled shearlet transform) coefficients belonging to the edge-related class remain unchanged, and those not belonging to the edge-related class are denoised by the shrinkage method using an adaptive threshold. Experimental results on various test images including benchmark grayscale images and medical ultrasound images demonstrate that the proposed method achieves better performance compared to some state-of-the-art denoising approaches.

Multi-Focus Image Fusion Based on Adaptive Dual-Channel Spiking Cortical Model in Non-Subsampled Shearlet Domain

To get a better fused performance in the multi-focus image fusion based on a transform domain, a new multi-focus image algorithm combined with the adaptive dual-channel spiking cortical model (SCM) in non-subsampled shearlet (NSST) domain and the difference images is proposed in this paper. First, a basic fused image is constructed in the NSST domain by registering the source image and adaptive dual channel SCM (dual-channel SCM). Next, the focus areas of the sources input images based on the difference images between the basic fused image and the sources images are detected. In the end, the final fused image generated in this paper is realized by combining the focal regions. Because of the global coupling of the dual-SCM, the synchronization characteristics of the pulse, and the multi-resolution and direction of the NSST, the proposed algorithm can preserve the information of the source’s image well and present a clear image more in line with the human visual effects. In summary, the image fusion algorithm that we have designed is superior to the most advanced algorithms.

Multimodal Neuroimaging Fusion in Nonsubsampled Shearlet Domain Using Location-Scale Distribution by Maximizing the High Frequency Subband Energy

A fusion of medical imaging data obtained from different modalities plays an important role in the current clinical practice. In this paper, we propose a novel multimodal fusion algorithm for brain imaging data based on the statistical properties of nonsubsampled shearlet transform (NSST) coefficients and a novel energy maximization fusion rule. The marginal distributions of the high-frequency NSST coefficients exhibit heavier tails than the Gaussian distribution. As a consequence, after studying its characteristics, we use a heavy-tailed probability density function, student’s $t$ location-scale distribution, to describe the highly non-Gaussian statistics of empirical NSST coefficients by learning the parameters using maximum likelihood estimation. Then, we employ this model to develop a maximum a posteriori estimator to obtain the noise-free coefficients. Then, for the first time, a novel fusion rule for obtaining the fused NSST coefficients based on maximizing the energy in the high-frequency subbands is proposed. Experiments are carried out on fusing two or more multimodal neuroimages taken from the BrainWeb, Alzheimer’s Disease Neuroimaging Initiative (ADNI), and Whole Brain Atlas databases. It is seen from the subjective and objective results that the proposed multimodal neuroimaging fusion method significantly outperforms the state-of-the-art methods including under noisy scenarios, and hence, it is more robust. It is also observed that the signal intensities in the fused images are better enhanced when a more number of source images are being fused. The proposed technique should benefit the medical professionals in diagnosing neurological disorders, such as Alzheimer, epilepsy, and multiple sclerosis.

Multi-Focus Image Fusion Based on Residual Network in Non-Subsampled Shearlet Domain

In order to obtain a panoramic image which is clearer, and has more layers and texture features, we propose an innovative multi-focus image fusion algorithm by combining with non-subsampled shearlet transform (NSST) and residual network (ResNet). First, NSST decomposes a pair of input images to produce subband coefficients of different frequencies for subsequent feature processing. Then, ResNet is applied to fuse the low frequency subband coefficients, and improved gradient sum of Laplace energy (IGSML) perform high frequency feature information processing. Finally, the inverse NSST is performed on the fused coefficients of different frequencies to obtain the final fused image. In our method, we fully consider the low frequency global features and high frequency detail information in image by using NSST. For low-frequency coefficients fusion, we can also obtain the spatial information features of low-frequency coefficient images by using ResNet, which has a deep network structure. IGSML can use different directional gradients to process high-frequency subband coefficients of different levels and directions, which is more conducive to the fusion of the coefficients. The experiment results show that the proposed method has been improved in the structural features and edge texture in the fusion images.

Deep feature learning for soft tissue sarcoma classification in MR images via transfer learning

Medical image analysis is motivated by deep learning emergence and computation power increase. Meanwhile, relevant deep features can significantly enhance learnable expert and intelligent systems performance and reduce diagnosis time and arduousness. This paper presents a deep learning-based radiomics framework for aided diagnosis of soft tissue sarcomas of the extremities. MR Images with histologically confirmed Liposarcoma (LPS) and Leiomyosarcomas (LMS) have been retrieved from the Cancer Imaging Archives database and pre-processed to recuperate ROIs from MR scans with delineated tumors. This study investigates the significance and impact of medical image fusion on deep feature learning based on transfer learning from the natural domain to the medical domain. Towards this end, we propose to fuse T1 with T2FS or STIR modalities using type-2 fuzzy sets in the non-subsampled shearlet domain. Being decomposed, low-frequency sub-images were selected using local energy and type-2 fuzzy entropy, while high frequencies were selected according to the maximum of the absolute value. Experimental results indicated that the proposed fusion framework outperformed the state-of-the-art fuzzy logic-based fusion techniques in terms of entropy and mutual information. Accordingly, we fine-tuned the pre-trained AlexNet deep convolutional neural network (CNN) with stochastic gradient descent (SGD). First, with the pre-processed dataset, and second with the fused images. As a result, the average classification accuracy using the augmented training data by image rotation and flipping was 97.17% with the raw data and 98.28% with the fused images, which highlighted the usefulness of complementary information for deep feature learning. One crucial concern was to investigate the depth of knowledge transferability. We incrementally fine-tuned the pre-trained CNN to assess the required level that achieves performance improvements in STS classification. Through layer-wise fine-tuning, our study further confirms the potential of middle and deep layers in performance improvement. Moreover, the transferability was concluded better than random weights. With the encouragement of classification results, our aided diagnosis framework may be in the pipeline to assist radiologists in classifying LPS and LMS.

Multimodal neurological image fusion based on adaptive biological inspired neural model in nonsubsampled Shearlet domain

AbstractMultimodal medical image fusion (MMIF) has a significant role for a better visualization of the diagnostic statistics, those help the medical professionals in the precise diagnosis of several critical diseases. This paper presents an improved fusion framework that uses the entire features extracted by the nonsubsampled shearlet transform (NSST) and adaptive biologically inspired neural model. The proposed scheme retains the required information without losing the resolution of the disease morphology. In the proposed method, the adaptive neural model based on local visibility and log Gabor energy based rules are applied to low and high‐frequency components, respectively. The better fusion results obtained by the proposed approach are confirmed by a large extent of simulations on the different MR‐SPECT and CT‐MR neurological images. Based on all these simulated results, it states that the proposed approach is superior than the other approaches as it produces better visually fused images with improved computational measures.

CT and MR image information fusion scheme using a cascaded framework in ripplet and NSST domain

The fusion of multimodal medical information is considered as an assisted approach for the medical professionals. Computed tomography and magnetic resonance (CT–MR) medical image fusion are able to help the radiologist in precise diagnosis of disease and deciding the required treatment in accord with the patient's condition. Therefore, a cascaded framework is proposed in this study that presents a fusion approach for multimodal medical information in ripplet transform (RT) and non-subsampled shearlet (NSST) domain. The RT and NSST having different features are utilised in a cascade manner that provides several directional decomposition coefficients and increases shift invariance information in the fused images. At the first stage decomposition, a biologically inspired neural model, motivated by novel sum-modified Laplacian and spatial frequency is utilised to fuse the low- and high-frequency coefficients, respectively, and the max fusion rule based on regional energy is utilised at stage 2. This model is used to preserve the redundant information also. The fusion performance is also validated by extensive simulations performed on different CT–MR image datasets of different diseases. Experimental results demonstrate that the proposed method provides better fused images in terms of visual quality along with the quantitative indices compared with several existing fusion approaches.

Nonsubsampled shearlet domain fusion techniques for CT–MR neurological images using improved biological inspired neural model

The fusion of multimodality medical images performs a very crucial role in the clinical diagnosis, analysis and the treatment of especially in critical diseases. It is considered as an assisted approach for the radiologist by providing the composite images having significant diagnostic information acquired from the source images. The main purpose of this work is to develop an efficient framework for fusing the multimodal medical images. Three different fusion techniques are proposed in this paper that presents the CT and MR medical image fusion in nonsubsampled shearlet transform (NSST) domain using the adaptive spiking neural model. The NSST having different features and a competent depiction of the image coefficients provides several directional decomposition coefficients. Maximum selection approach and regional energy are utilized for low frequency coefficients fusion. Spatial frequency, novel modified spatial frequency and novel sum modified Laplacian motivated spiking model are used for every high frequency subimage component. Finally, fused images are reconstructed by applying inverse NSST. The performance of proposed fusion techniques is validated by extensive simulations performed on different CT-MR image datasets using proposed and other thirty seven existing fusion approaches in terms of both the subjective and objective manner. The results revealed that the proposed techniques provide better visualization of resultant images and higher quantitative measures compared to several existing fusion approaches.

Blind optimum detector for robust image watermarking in nonsubsampled shearlet Domain

In this paper, we propose a blind robust image watermarking approach in the nonsubsampled Shearlet domain using Bessel K Form (BKF) modeling. Nonsubsampled shearlet transform (NSST) is an effective multi-scale and multi-direction analysis method, it not only can exactly compute the shearlet coefficients based on a multiresolution analysis, but also can provide nearly optimal approximation for a piecewise smooth function. In order to achieve more robustness and imperceptibility, watermark information is embedded into the most significant NSST directional subband with the highest energy, by simply modifying the NSST coefficient amplitude according to the watermark data. By modeling the NSST coefficients with BKF probability density function, the distribution of watermarked noisy coefficients is analytically calculated. The tradeoff between the imperceptibility and robustness of watermark data is solved in a novel fashion. At the receiver, based on the Maximum Likelihood (ML) decision rule, a blind statistical watermark detector is proposed. Experimental results on test images demonstrate that the proposed approach can provide better imperceptibility and robustness against various attacks, such as additive white Gaussian noise, salt & pepper noise, median filtering, JPEG compression, rotation, and scaling, in comparison with the recently proposed techniques.

Despeckling of ultrasound medical images using nonlinear adaptive anisotropic diffusion in nonsubsampled shearlet domain

Despeckling is of great interest in ultrasound medical images. The inherent limitations of acquisition techniques and systems introduce the speckles in ultrasound images. These speckles are the main factors that degrade the quality and most importantly texture information present in ultrasound images. Due to these speckles, experts may not be able to extract correct and useful information from the images. This paper presents an edge preserved despeckling approach that combines the nonsubsampled shearlet transform (NSST) with improved nonlinear diffusion equations. As a new image representation method with the different features of localization, directionality and multiscale, the NSST is utilized to provide the effective representation of the image coefficients. The anisotropic diffusion approach is applied to the noisy coarser NSST coefficients to improve the noise reduction efficiency and effectively preserves the edge features. In the diffusion process, an adaptive gray variance is also incorporated with the gradient information of eight connected neighboring pixels to preserve the edges, effectively. The performance of the proposed method is evaluated by conducting extensive simulations using both the standard test images and several ultrasound medical images. Experiments show that the proposed method provides an improvement not only in noise reduction but also in the preservation of more edges as compared to several existing methods.

Image denoising by non-subsampled shearlet domain multivariate model and its method noise thresholding

The denoising of a natural image corrupted by additive Gaussian white noise is a classical problem in image processing. A new image denoising method is proposed by using the dependencies between the non-subsampled shearlet transform (NSST) coefficients and their neighbors. The NSST is well known for its approximate shift invariance and better directional selectivity, which are very important in image denoising. In order to take account of the dependency between the current NSST coefficient and its neighbors, a new multivariate probability density function (pdf) is exploited, and a multivariate shrinkage function for image denoising is derived from it by using the maximum a posteriori (MAP) estimator. Then, the blend of the proposed multivariate shrinkage function and its method noise thresholding using NSST is proposed for image denoising. Experimental results demonstrate that the proposed method is very competitive when compared with other existing denoising methods in the literature.