Multi-modal medical image fusion via multi-dictionary and truncated Huber filtering

Biomedical Signal Processing and Control, Год журнала: 2023, Номер 88, С. 105671 - 105671

Опубликована: Ноя. 4, 2023

Язык: Английский

---

Analysis of Multimodality Fusion of Medical Image Segmentation Employing Deep Learning

Опубликована: Фев. 19, 2024

Язык: Английский

---

Deep-Learning Based Multi-Modalities Fusion for the Detection of Brain-Related Diseases: A Review

Communications in computer and information science, Год журнала: 2025, Номер unknown, С. 149 - 170

Опубликована: Янв. 1, 2025

Язык: Английский

---

A rapid multi-parametric quantitative MR imaging method to assess Parkinson’s disease: a feasibility study

BMC Medical Imaging, Год журнала: 2024, Номер 24(1)

Опубликована: Март 5, 2024

Abstract Background MULTIPLEX is a single-scan three-dimensional multi-parametric MRI technique that provides 1 mm isotropic T1-, T2*-, proton density- and susceptibility-weighted images the corresponding quantitative maps. This study aimed to investigate its feasibility of clinical application in Parkinson’s disease (PD). Methods 27 PD patients 23 healthy control (HC) were recruited underwent scanning. All image reconstruction processing automatically performed with in-house C + programs on Automatic Differentiation using Expression Template platform. According HybraPD atlas consisting 12 human brain subcortical nuclei, region-of-interest (ROI) based analysis was conducted extract parameters, then identify PD-related abnormalities from T1, T2* density maps susceptibility mapping (QSM), by comparing HCs. Results The ROI-based revealed significantly decreased mean T1 values substantia nigra pars compacta habenular value subthalamic nucleus increased QSM patients, compared HCs (all p < 0.05 after FDR correction). receiver operating characteristic showed all these four parameters contributed diagnosis 0.01 Furthermore, two hemicerebral differences regard clinically dominant side among patients. Conclusions might be feasible for assist provide possible pathological information patients’ dopaminergic midbrain regions.

Язык: Английский

---

An ensemble deep learning model for medical image fusion with Siamese neural networks and VGG-19

PLoS ONE, Год журнала: 2024, Номер 19(10), С. e0309651 - e0309651

Опубликована: Окт. 23, 2024

Multimodal medical image fusion methods, which combine complementary information from many multi-modality images, are among the most important and practical approaches in numerous clinical applications. Various conventional techniques have been developed for multimodality fusion. Complex procedures weight map computing, fixed strategy lack of contextual understanding remain difficult machine learning approaches, usually resulting artefacts that degrade quality. This work proposes an efficient hybrid model using pre-trained non-pre-trained networks i.e. VGG-19 SNN with stacking ensemble method. The leveraging unique capabilities each architecture, can effectively preserve detailed high visual quality, combinations modalities challenges, notably improved contrast, increased resolution, lower artefacts. Additionally, this be more robust various source images publicly available Havard-Medical-Image-Fusion Datasets, GitHub. Kaggle. Our proposed performance is superior terms quality metrics to existing methods literature like PCA+DTCWT, NSCT, DWT, DTCWT+NSCT, GADCT, CNN VGG-19.

Язык: Английский

---

Tensor Methods in Biomedical Image Analysis

Journal of Medical Signals & Sensors, Год журнала: 2024, Номер 14(6)

Опубликована: Июнь 1, 2024

Abstract In the past decade, tensors have become increasingly attractive in different aspects of signal and image processing areas. The main reason is inefficiency matrices representing analyzing multimodal multidimensional datasets. Matrices cannot preserve correlation elements higher-order datasets this highly reduces effectiveness matrix-based approaches Besides this, tensor-based demonstrated promising performances. These together, encouraged researchers to move from tensors. Among applications, biomedical signals images particular importance. This due need for extracting accurate information which directly affects patient’s health. addition, many cases, several been recorded simultaneously a patient. A common example recording electroencephalography (EEG) functional magnetic resonance imaging (fMRI) patient with schizophrenia. such situation, seem be among most effective methods simultaneous exploitation two (or more) Therefore, developed Considering reality, paper, we aim comprehensive review on analysis. presented study classification between applications can show importance enhancement open new ways future studies.

Язык: Английский

---

MIF-BTF-MRN: Medical image fusion based on the bilateral texture filter and transfer learning with the ResNet-101 network

Biomedical Signal Processing and Control, Год журнала: 2024, Номер 100, С. 106976 - 106976

Опубликована: Окт. 10, 2024

Язык: Английский

---

Comprehensive performance analysis of different medical image fusion techniques for accurate healthcare diagnosis applications

Multimedia Tools and Applications, Год журнала: 2023, Номер 83(8), С. 24217 - 24276

Опубликована: Авг. 10, 2023

Язык: Английский

---

SRDRN-IR: A Super Resolution Deep Residual Neural Network for IR Images

2022 IEEE World Conference on Applied Intelligence and Computing (AIC), Год журнала: 2023, Номер unknown, С. 746 - 751

Опубликована: Июль 29, 2023

Human perception is only capable of perceiving a few objects outside the range wavelengths for visible light in electromagnetic spectrum. It restricts humans' ability to discriminate things variety situations, such as dim or under smoke and fog. The development thermographic imaging technology has made it possible see items that are invisible naked eye. This enables its usage many sectors, defence, agriculture, healthcare, etc. Thermal cameras have low spatial resolution comparison same-range RGB due hardware constraints. A deep neural network architecture, SRDRN, proposed this study Super-Resolution (SR) IR images. SRDRN uses channel splitting concept with residual learning computationally efficient super resolution. viability design validated by analysing available thermal image datasets.

Язык: Английский

---

A Multi-Modal Framework for Fake News Analysis for Detection to Deconstruction

Опубликована: Март 15, 2024

The research discusses a multimodal framework for analyzing and detecting fake news understanding its impact on society. This employs diverse strategies, including linguistic analysis, social network monitoring, visual assessment, to capture various aspects of fabricated information dissemination. first component the focuses examining language textual content used in articles identify misleading information, biased language, sensationalized headlines. second analyzes role networks spreading news, tracking propagation through platforms like media identifying key actors influencers involved third involves images videos manipulate public perceptions emotions, detection doctored illustrations.

Язык: Английский

---