Enhancing Ulcerative Colitis Diagnosis: A Multi-Level Classification Approach with Deep Learning

Computer Modeling in Engineering & Sciences, Год журнала: 2024, Номер 140(1), С. 1129 - 1142

Опубликована: Янв. 1, 2024

The evaluation of disease severity through endoscopy is pivotal in managing patients with ulcerative colitis, a condition significant clinical implications.However, endoscopic assessment susceptible to inherent variations, both within and between observers, compromising the reliability individual evaluations.This study addresses this challenge by harnessing deep learning develop robust model capable discerning discrete levels severity.To initiate endeavor, multi-faceted approach embarked upon.The dataset meticulously preprocessed, enhancing quality discriminative features images contrast limited adaptive histogram equalization (CLAHE).A diverse array data augmentation techniques, encompassing various geometric transformations, leveraged fortify dataset's diversity facilitate effective feature extraction.A fundamental aspect involves strategic incorporation transfer principles, modified ResNet-50 architecture.This augmentation, informed domain expertise, contributed significantly model's classification performance.The outcome research endeavor yielded highly promising model, demonstrating an accuracy rate 86.85%, coupled recall 82.11% precision 89.23%.

Язык: Английский

---

Vision Transformer Model for Gastrointestinal Tract Diseases Classification from WCE Images

Опубликована: Апрель 21, 2024

Accurate disease classification utilizing endoscopic images indeed poses a significant challenge within the field of gastroenterology. This research introduces methodology for assisting medical diagnostic procedures and detecting gastrointestinal (GI) tract diseases by categorizing features extracted from using Vision Transformer (ViT) models. We propose three ViT-inspired models classifying GI colon acquired through wireless capsule endoscopy (WCE). The highest achieved accuracy among our is 97.83%. conducted comparative analysis with pre-trained CNN (Convolutional Neural Network) namely, Xception, DenseNet121, MobileNet, alongside recent papers to validate findings.

Язык: Английский

---

Explainable AI for gastrointestinal disease diagnosis in telesurgery Healthcare 4.0

Computers & Electrical Engineering, Год журнала: 2024, Номер 118, С. 109414 - 109414

Опубликована: Июль 8, 2024

Язык: Английский

---

Applications of Artificial Intelligence and Machine Learning in Spine MRI

Bioengineering, Год журнала: 2024, Номер 11(9), С. 894 - 894

Опубликована: Сен. 5, 2024

Diagnostic imaging, particularly MRI, plays a key role in the evaluation of many spine pathologies. Recent progress artificial intelligence and its subset, machine learning, has led to applications within which we sought examine this review. A literature search major databases (PubMed, MEDLINE, Web Science, ClinicalTrials.gov) was conducted according Preferred Reporting Items for Systematic Reviews Meta-Analyses (PRISMA) guidelines. The yielded 1226 results, 50 studies were selected inclusion. Key data from these extracted. Studies categorized thematically into following: Image Acquisition Processing, Segmentation, Diagnosis Treatment Planning, Patient Selection Prognostication. Gaps proposed areas future research are discussed. Current demonstrates ability improve various aspects field, image acquisition analysis clinical care. We also acknowledge limitations current technology. Future work will require collaborative efforts order fully exploit new technologies while addressing practical challenges generalizability implementation. In particular, use foundation models large-language MRI is promising area, warranting further research. assessing model performance real-world settings help uncover unintended consequences maximize benefits patient

Язык: Английский

---

Enhanced Skin Disease Classification via Dataset Refinement and Attention-Based Vision Approach

Bioengineering, Год журнала: 2025, Номер 12(3), С. 275 - 275

Опубликована: Март 11, 2025

Skin diseases are listed among the most frequently encountered diseases. such as eczema, melanoma, and others necessitate early diagnosis to avoid further complications. This study aims enhance of skin disease by utilizing advanced image processing techniques an attention-based vision approach support dermatologists in solving classification problems. Initially, is being passed through various steps quality dataset. These adaptive histogram equalization, binary cross-entropy with implicit averaging, gamma correction, contrast stretching. Afterwards, enhanced images for performing which based on encoder part transformers multi-head attention. Extensive experimentation performed collect results two publicly available datasets show robustness proposed approach. The evaluation shows competitive compared a state-of-the-art

Язык: Английский

---

A novel network-level fused deep learning architecture with shallow neural network classifier for gastrointestinal cancer classification from wireless capsule endoscopy images

BMC Medical Informatics and Decision Making, Год журнала: 2025, Номер 25(1)

Опубликована: Март 31, 2025

Deep learning has significantly contributed to medical imaging and computer-aided diagnosis (CAD), providing accurate disease classification diagnosis. However, challenges such as inter- intra-class similarities, class imbalance, computational inefficiencies due numerous hyperparameters persist. This study aims address these by presenting a novel deep-learning framework for classifying localizing gastrointestinal (GI) diseases from wireless capsule endoscopy (WCE) images. The proposed begins with dataset augmentation enhance training robustness. Two architectures, Sparse Convolutional DenseNet201 Self-Attention (SC-DSAN) CNN-GRU, are fused at the network level using depth concatenation layer, avoiding costs of feature-level fusion. Bayesian Optimization (BO) is employed dynamic hyperparameter tuning, an Entropy-controlled Marine Predators Algorithm (EMPA) selects optimal features. These features classified Shallow Wide Neural Network (SWNN) traditional classifiers. Experimental evaluations on Kvasir-V1 Kvasir-V2 datasets demonstrate superior performance, achieving accuracies 99.60% 95.10%, respectively. offers improved accuracy, precision, efficiency compared state-of-the-art models. addresses key in GI diagnosis, demonstrating its potential efficient clinical applications. Future work will explore adaptability additional optimize complexity broader deployment.

Язык: Английский

---

Deep ensemble learning for gastrointestinal diagnosis using endoscopic image classification

PeerJ Computer Science, Год журнала: 2025, Номер 11, С. e2809 - e2809

Опубликована: Апрель 22, 2025

Transfer learning is a valuable tool for the effective assistance of gastroenterologists in powerful diagnosis medical images with fast convergence. It also intends to minimize time and estimated effort required improved gastrointestinal tract (GIT) diagnosis. GIT abnormalities are widely known be fatal disorders leading significant mortalities. includes both upper lower disorders. The challenges addressing issues complex need study. Multiple exist regarding computer-aided (CAD) endoscopy including lack annotated images, dark backgrounds, less contrast, noisy irregular patterns. Deep transfer have assisted various ways. goal proposed framework classification endoscopic enhanced accuracy. research aims formulate learning-based deep ensemble model, accurately classifying therapeutic purposes. model based on weighted voting two state-of-the-art (STA) base models, NasNet-Mobile EfficientNet. extraction regions interest, specifically sick portions, been performed using captured from procedure. Performance evaluation cross-dataset evaluation. datasets utilized include training dataset HyperKvasir test datasets, Kvasir v1 v2. However, alone cannot create robust due unequal distribution across categories, making promising approach development. has conducted by utilizing accuracy, precision, recall, Area under curve (AUC) score F1 performance metrics. work outperforms much existing models giving 97.83% 98.45% accuracy

Язык: Английский

---

SNet: A novel convolutional neural network architecture for advanced endoscopic image classification of gastrointestinal disorders

SLAS TECHNOLOGY, Год журнала: 2025, Номер unknown, С. 100304 - 100304

Опубликована: Май 1, 2025

With the intent of assisting gastroenterologists from all over world, proposed work aims to eliminate effort required achieve accurate diagnoses. Statistically, gastrointestinal diseases often result in fatal disorders, contributing a significant number fatalities. The upper tract (GIT) includes stomach, esophagus, and duodenum, while lower one comprises section small intestine, namely ileum, as well large including colon. challenges associated with GIT issues are apparently complex. Therefore, multiple exist regarding CAD (Computer-aided diagnosis) endoscopy, lack annotated images, dark background, poor contrast, an irregular pattern. objective this research is develop robust deep network, called SNet, that offers solution complex classification problems. Firstly, endoscopic images undergo preprocessing before being subjected feature extraction. This step involves image resizing along augmentation step. convolutional neural network (CNN) model comprised six blocks placed at different layers. To enable exhaustive evaluation framework across datasets, has undergone training on very HyperKvasir dataset, later tested Kvasir v1 v2 datasets. facilitates cross-dataset system evaluation, resulting efficient for unseen diagnosis. avoid problem "curse dimensionality", most discriminant information selected based minimum redundancy maximum relevance (MRMR) algorithm. architecture been evaluated using range performance metrics, such accuracy, sensitivity, specificity, Area under curve (AUC). accuracy main metric, achieved 98.45% v1, 97.83%

Язык: Английский

---

Analysis of Colorectal and Gastric Cancer Classification: A Mathematical Insight Utilizing Traditional Machine Learning Classifiers

Mathematics, Год журнала: 2023, Номер 11(24), С. 4937 - 4937

Опубликована: Дек. 12, 2023

Cancer remains a formidable global health challenge, claiming millions of lives annually. Timely and accurate cancer diagnosis is imperative. While numerous reviews have explored classification using machine learning deep techniques, scant literature focuses on traditional ML methods. In this manuscript, we undertake comprehensive review colorectal gastric detection specifically employing classifiers. This emphasizes the mathematical underpinnings detection, encompassing preprocessing feature extraction, classifiers, performance assessment metrics. We provide formulations for these key components. Our analysis limited to peer-reviewed articles published between 2017 2023, exclusively considering medical imaging datasets. Benchmark publicly available datasets cancers are presented. synthesizes findings from 20 16 cancer, culminating in total 36 research articles. A significant focus placed commonly used features, Crucially, introduce our optimized methodology both cancers. metrics reveals remarkable results: 100% accuracy types, but with lowest sensitivity recorded at 43.1% cancer.

Язык: Английский

---

Advancements in traditional machine learning techniques for detection and diagnosis of fatal cancer types: Comprehensive review of biomedical imaging datasets

Measurement, Год журнала: 2023, Номер 225, С. 114059 - 114059

Опубликована: Дек. 22, 2023

Язык: Английский

---