A Review of Datasets, Optimization Strategies, and Learning Algorithms for Analyzing Alzheimer’s Dementia Detection DOI Creative Commons

Vanaja Thulasimani,

Kogilavani Shanmugavadivel, Jaehyuk Cho

и другие.

Neuropsychiatric Disease and Treatment, Год журнала: 2024, Номер Volume 20, С. 2203 - 2225

Опубликована: Ноя. 1, 2024

Alzheimer's Dementia (AD) is a progressive neurological disorder that affects memory and cognitive function, necessitating early detection for its effective management. This poses significant challenge to global public health. The accurate of dementia crucial several reasons. First, timely facilitates intervention planning treatment. Second, precise diagnostic methods are essential distinguishing from other disorders medical conditions may present with similar symptoms. Continuous analysis improvements in have contributed advancements research. It helps identify new biomarkers, refine existing tools, foster the development innovative technologies, ultimately leading more efficient approaches dementia. paper presents critical multimodal imaging datasets, learning algorithms, optimisation techniques utilised context detection. focus on understanding challenges employing diverse modalities, such as MRI (Magnetic Resonance Imaging), PET (Positron Emission Tomography), EEG (ElectroEncephaloGram). study evaluated various machine deep models, transfer techniques, generative adversarial networks multi-modality data In addition, examination encompassing algorithms hyperparameter tuning strategies processing analysing images presented this discern their influence model performance generalisation. Thorough enhancement fundamental addressing healthcare posed by dementia, facilitating interventions, improving accuracy, advancing research neurodegenerative diseases.

Язык: Английский

XTNSR: Xception-based transformer network for single image super resolution DOI Creative Commons
Jagrati Talreja, Supavadee Aramvith, Takao Onoye

и другие.

Complex & Intelligent Systems, Год журнала: 2025, Номер 11(2)

Опубликована: Янв. 25, 2025

Язык: Английский

Процитировано

2
A new superfluity deep learning model for detecting knee osteoporosis and osteopenia in X-ray images DOI Creative Commons
Soaad M. Naguib, M. Saleh, Hanaa M. Hamza

и другие.

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Окт. 25, 2024

This study proposes a new deep-learning approach incorporating superfluity mechanism to categorize knee X-ray images into osteoporosis, osteopenia, and normal classes. The suggests the use of two distinct types blocks. rationale is that, unlike conventional serially stacked layer, concept involves concatenating multiple layers, enabling features flow branches rather than single branch. Two datasets have been utilized for training, validating, testing proposed model. We transfer learning with pre-trained models, AlexNet ResNet50, comparing results those indicate that performance namely was inferior Superfluity DL architecture. model demonstrated highest accuracy (85.42% dataset1 79.39% dataset2) among all models.

Язык: Английский

Процитировано

4
Improved sarcoidosis disease detection using deep learning and histogram of oriented gradients with quantum SVM DOI Creative Commons
Aleka Melese Ayalew, Worku Abebe Degife, Nigus Wereta Asnake

и другие.

Deleted Journal, Год журнала: 2025, Номер 7(3)

Опубликована: Фев. 19, 2025

Язык: Английский

Процитировано

0
Attention LinkNet-152: a novel encoder-decoder based deep learning network for automated spine segmentation DOI Creative Commons

Aqsa Dastgir,

Bin Wang, Muhammad Usman Saeed

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Апрель 16, 2025

Abstract Segmenting the spine from CT images is crucial for diagnosing and treating spine-related conditions but remains challenging due to spine’s complex anatomy imaging artifacts. This study introduces a novel encoder-decoder-based deep learning approach, named LinkNet-152, tailored automated segmentation. The model integrates modified EfficientNetB7 encoder with attention modules enhance feature extraction by focusing on regions of interest. decoder leverages LinkNet architecture, replacing ResNet34 deeper ResNet152 improve segmentation accuracy. Additionally, gradient sensitivity-based pruning applied optimize model’s complexity computational efficiency. Evaluated VerSe 2019 2020 datasets, proposed achieves superior performance, Dice coefficient 96.85% Jaccard index 95.37%, outperforming state-of-the-art methods. These results highlight effectiveness in addressing challenges its potential advance clinical applications.

Язык: Английский

Процитировано

0
Diagnosis of Paddy Diseases Using Pre-Trained Architectures and a Proposed Enhanced EfficientNetB3 Model DOI

B Johnson,

Chandrakumar Thangavel

Tarım Bilimleri Dergisi, Год журнала: 2025, Номер 31(2), С. 558 - 576

Опубликована: Март 25, 2025

Rice is an important crop in India and often affected by pests diseases, which can lead to a significant drop production. This research investigates advanced deep learning approaches for accurate paddy disease diagnosis, focusing on comparing several transfer models. The study specifically targets diseases such as Tungro, Dead Heart, Hispa, Blast, Downy Mildew, Brown Spot, Bacterial Leaf Blight, Panicle Streak. base EfficientNetB3 model attains approximately 95.55 % accuracy during training 95.12% evaluation unseen data. However, it encounters challenges when applied domain-specific tasks diagnosing frequently experiencing issues overfitting inadequate convergence. To overcome these issues, Enhanced was developed, incorporating batch normalization, dropout, data regularization techniques. conducted using the 'Paddy Doctor' dataset, featuring 10,407 high- resolution images of leaves. It reached 98.92 with loss rate 0.1385. For validation, 98.20 0.1450. On independent test set, 98.50 obtained 0.1505. With remarkable time just 68 minutes, demonstrates its potential precise diagnosis. Its impressive performance plays crucial role advancing management boosting yields.

Язык: Английский

Процитировано

0
Attention-Based Dual-Path Deep Learning for Blood Cell Image Classification Using ConvNeXt and Swin Transformer DOI
Şafak Kılıç

Deleted Journal, Год журнала: 2025, Номер unknown

Опубликована: Апрель 29, 2025

In the rapidly evolving field of medical image analysis, precise classification blood cells plays a crucial role in diagnosing and monitoring numerous hematological disorders. Traditional methods, while effective, often require significant manual effort expert knowledge, leading to potential delays inconsistencies diagnosis. Addressing these challenges, this paper introduces groundbreaking dual-path deep learning architecture that synergistically combines ConvNeXt Swin Transformer networks. This innovative approach leverages strengths convolutional neural networks for local feature extraction transformers global context integration, effectively capturing complex morphological variations cells. Furthermore, incorporation Multi-scale Preprocessing Module (MPM) significantly enhances quality, employing techniques such as contrast enhancement, illumination normalization, enhancement improve visibility differentiation cellular features. Tested on comprehensive dataset 17,092 cell images, our model achieves an unprecedented accuracy 99.98%, demonstrating superior performance over existing methods. level not only underscores effectiveness but also highlights its serve reliable tool clinical settings, facilitating faster more accurate analysis. By automating process with high precision, promises enhance diagnostic workflows, reduce workload professionals, ultimately contribute better patient outcomes hematology.

Язык: Английский

Процитировано

0
Classification of the stages of Alzheimer’s disease based on three-dimensional lightweight neural networks DOI Creative Commons
Jun Li, Juntong Liu, Su Yang

и другие.

PeerJ Computer Science, Год журнала: 2025, Номер 11, С. e2897 - e2897

Опубликована: Май 15, 2025

Alzheimer’s disease is a neurodegenerative that seriously threatens the life and health of elderly. This study used three-dimensional lightweight neural networks to classify stages explore relationship between variations brain tissue. The CAT12 preprocess magnetic resonance images got three kinds preprocessed images: standardized images, segmented gray matter white images. were train four respectively, evaluation metrics are calculated. accuracies for classifying (cognitively normal, mild cognitive impairment, disease) in above 96%, precisions recalls 94%. found classification cognitively best results can be obtained by training with impairment disease, analyzed process normal more obvious at beginning, while not obvious. As progresses, tend become significant, both significant development disease.

Язык: Английский

Процитировано

0
Modified MobileNet with leaky ReLU and LSTM with balancing technique to classify the soil types DOI
Kamini G. Panchbhai, Madhusudan G. Lanjewar, Aditi Venkatesh Naik

и другие.

Earth Science Informatics, Год журнала: 2024, Номер 18(1)

Опубликована: Дек. 19, 2024

Язык: Английский

Процитировано

1
Effect of Data Augmentation Method in Applied Science Data-Based Salt Area Estimation with U-Net DOI
Betül Ağaoğlu, I. N. Askerzade, Erkan Bostancı

и другие.

Türkiye teknoloji ve uygulamalı bilimler dergisi., Год журнала: 2024, Номер 5(2), С. 70 - 86

Опубликована: Окт. 5, 2024

Oil and natural gas rank first as energy inputs worldwide. Other subsurface resources, such salt, provide clues to obtaining these resources. Salt accumulation areas are resources used locate oil fields. Seismic images, which geological data, information for locating underground Manual interpretation of images requires expert knowledge experience. This time-consuming laborious method is also limited by the fact that it cannot be replicated. Deep learning a very successful image segmentation in recent years. Automating detection reserves seismic using artificial intelligence methods reduces time, cost workload factors. In this study, we aim identify salt U-net architecture on identification challenge shared TGS (the world’s leading geoscience data company) Identification Challenge kaggle.com. addition, effect augmentation designed system investigated. The set consists combined together automatic mass. study aims obtain highest accuracy lowest error rate detect from images. As result IoU (Intersection over Union) value without 0.9390, while 0.9445.

Язык: Английский

Процитировано

0
Deep Learning for Automatic Classification of Fruits and Vegetables: Evaluation from the Perspectives of Efficiency and Accuracy DOI
Demet Parlak Sönmez, Şafak Kılıç

Türkiye teknoloji ve uygulamalı bilimler dergisi., Год журнала: 2024, Номер 5(2), С. 151 - 171

Опубликована: Окт. 5, 2024

Within the agricultural domain, accurately categorizing freshness levels of fruits and vegetables holds immense significance, as this classification enables early detection spoilage allows for appropriate grouping products based on their intended export destinations. These processes necessitate a system capable meticulously classifying while minimizing labor expenditures. The current study concentrates developing an advanced model that can effectively categorize status each fruit vegetable 'good,' 'medium,' or 'spoiled.' To achieve objective, various artificial intelligence models, including CNN, AlexNet, ResNet50, GoogleNet, VGG16, EfficientB3, have been implemented, attaining remarkable success rates 99.75%, 97.97%, 96.71%, 99.49%, 98.75%, 99.81%, respectively.

Язык: Английский

Процитировано

0