Automated Detection of Gastrointestinal Diseases Using Resnet50*-Based Explainable Deep Feature Engineering Model with Endoscopy Images DOI Creative Commons

Veysel Yusuf Cambay,

Prabal Datta Barua, Abdul Hafeez‐Baig

и другие.

Sensors, Год журнала: 2024, Номер 24(23), С. 7710 - 7710

Опубликована: Дек. 2, 2024

This work aims to develop a novel convolutional neural network (CNN) named ResNet50* detect various gastrointestinal diseases using new ResNet50*-based deep feature engineering model with endoscopy images. The novelty of this is the development ResNet50*, variant ResNet model, featuring convolution-based residual blocks and pooling-based attention mechanism similar PoolFormer. Using image dataset was trained, an explainable (DFE) developed. DFE comprises four primary stages: (i) extraction, (ii) iterative selection, (iii) classification shallow classifiers, (iv) information fusion. self-organizing, producing 14 different outcomes (8 classifier-specific 6 voted) selecting most effective result as final decision. During heatmaps are identified gradient-weighted class activation mapping (Grad-CAM) features derived from these regions via global average pooling layer pretrained ResNet50*. Four selectors employed in selection stage obtain distinct vectors. classifiers k-nearest neighbors (kNN) support vector machine (SVM) used produce specific outcomes. Iterative majority voting voted top determined by greedy algorithm based on accuracy. presented trained augmented version Kvasir dataset, its performance tested Kvasir, 2, wireless capsule (WCE) curated colon disease datasets. Our proposed demonstrated accuracy more than 92% for all three datasets remarkable 99.13% WCE dataset. These findings affirm superior ability confirm generalizability developed architecture, showing consistent across

Язык: Английский

Improving image classification of gastrointestinal endoscopy using curriculum self-supervised learning DOI Creative Commons

Han Guo,

Sai Ashish Somayajula,

Ramtin Hosseini

и другие.

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

Опубликована: Март 13, 2024

Abstract Endoscopy, a widely used medical procedure for examining the gastrointestinal (GI) tract to detect potential disorders, poses challenges in manual diagnosis due non-specific symptoms and difficulties accessing affected areas. While supervised machine learning models have proven effective assisting clinical of GI scarcity image-label pairs created by experts limits their availability. To address these limitations, we propose curriculum self-supervised framework inspired human learning. Our approach leverages HyperKvasir dataset, which comprises 100k unlabeled images pre-training 10k labeled fine-tuning. By adopting our proposed method, achieved an impressive top-1 accuracy 88.92% F1 score 73.39%. This represents 2.1% increase over vanilla SimSiam 1.9% score. The combination curriculum-based demonstrates efficacy advancing disorders. study highlights utilizing improve paving way more accurate efficient endoscopy.

Язык: Английский

Процитировано

3
Design and implementation of an AI-controlled spraying drone for agricultural applications using advanced image preprocessing techniques DOI
Cemalettin Akdoğan, Tolga Özer, Yüksel Oğuz

и другие.

Robotic Intelligence and Automation, Год журнала: 2024, Номер 44(1), С. 131 - 151

Опубликована: Март 18, 2024

Purpose Nowadays, food problems are likely to arise because of the increasing global population and decreasing arable land. Therefore, it is necessary increase yield agricultural products. Pesticides can be used improve land This study aims make spraying cherry trees more effective efficient with designed artificial intelligence (AI)-based unmanned aerial vehicle (UAV). Design/methodology/approach Two approaches have been adopted for AI-based detection trees: In approach 1, YOLOv5, YOLOv7 YOLOv8 models trained 70, 100 150 epochs. Approach 2, a new method proposed performance metrics obtained in 1. Gaussian, wavelet transform (WT) Histogram Equalization (HE) preprocessing techniques were applied generated data set 2. The best-performing 1 2 real-time test application developed UAV. Findings best F1 score was 98% epochs YOLOv5s model. mAP values as 98.6% 98.9% epochs, YOLOv5m model an improvement 0.6% score. tests, drone system detected sprayed accuracy 66% 77% It revealed that use pesticides could reduced by 53% energy consumption 47%. Originality/value An original created designing detect spray using AI. classify trees. results compared. including HE, Gaussian WT proposed, improved. effect experimental thoroughly analyzed.

Язык: Английский

Процитировано

3
Adaptive Treatment Planning via Multi-class Segmentation of GI Tract Tumours DOI

Samyak Jain,

Achira Pal,

J. Andrew

и другие.

Lecture notes in networks and systems, Год журнала: 2025, Номер unknown, С. 477 - 492

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
Advanced Hybridization and Optimization of DNNs for Medical Imaging: A Survey on Disease Detection Techniques DOI Creative Commons

Maneet Kaur Bohmrah,

Harjot Kaur

Artificial Intelligence Review, Год журнала: 2025, Номер 58(4)

Опубликована: Фев. 4, 2025

Язык: Английский

Процитировано

0
Vision Transformer-Based Classification of Gastrointestinal Ulcers Using WCE Images DOI
Srijita Bandopadhyay,

Joydeep Roy Chowdhury,

Rahul Shaw

и другие.

Lecture notes in networks and systems, Год журнала: 2025, Номер unknown, С. 367 - 378

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
The effect of wavelet transform on the classification performance of different deep learning architectures DOI
Muhammed Mustafa Kelek, Uğur Fidan, Yüksel Oğuz

и другие.

Signal Image and Video Processing, Год журнала: 2025, Номер 19(5)

Опубликована: Март 6, 2025

Язык: Английский

Процитировано

0
MultiResFF‐Net: Multilevel Residual Block‐Based Lightweight Feature Fused Network With Attention for Gastrointestinal Disease Diagnosis DOI Creative Commons
Sohaib Asif,

Yajun Ying,

Tingting Qian

и другие.

International Journal of Intelligent Systems, Год журнала: 2025, Номер 2025(1)

Опубликована: Янв. 1, 2025

Accurate detection of gastrointestinal (GI) diseases is crucial due to their high prevalence. Screening often inefficient with existing methods, and the complexity medical images challenges single‐model approaches. Leveraging diverse model features can improve accuracy simplify detection. In this study, we introduce a novel deep learning tailored for diagnosis GI through analysis endoscopy images. This innovative model, named MultiResFF‐Net, employs multilevel residual block‐based feature fusion network. The key strategy involves integration from truncated DenseNet121 MobileNet architectures. not only optimizes model’s diagnostic performance but also strategically minimizes computational demands, making MultiResFF‐Net valuable tool efficient accurate disease in A pivotal component enhancing introduction Modified MultiRes‐Block (MMRes‐Block) Convolutional Block Attention Module (CBAM). MMRes‐Block, customized component, optimally handles fused at endpoint both models, fostering richer sets without escalating parameters. Simultaneously, CBAM ensures dynamic recalibration maps, emphasizing relevant channels spatial locations. dual incorporation significantly reduces overfitting, augments precision, refines extraction process. Extensive evaluations on three datasets—endoscopic images, GastroVision data, histopathological images—demonstrate exceptional 99.37%, 97.47%, 99.80%, respectively. Notably, achieves superior efficiency, requiring 2.22 MFLOPS 0.47 million parameters, outperforming state‐of‐the‐art models cost‐effectiveness. These results establish as robust practical

Язык: Английский

Процитировано

0
Multi-parameter prediction of oil palm fruit quality through near infrared spectroscopy combined with chemometric analysis DOI
Muhammad Achirul Nanda, Kharistya Amaru, S. Rosalinda

и другие.

Spectrochimica Acta Part A Molecular and Biomolecular Spectroscopy, Год журнала: 2025, Номер 343, С. 126505 - 126505

Опубликована: Май 31, 2025

Язык: Английский

Процитировано

0
Color-Transfer-Enhanced Data Construction and Validation for Deep Learning-Based Upper Gastrointestinal Landmark Classification in Wireless Capsule Endoscopy DOI Creative Commons

Hae-tae Kim,

Byungwoo Cho, Jong‐Oh Park

и другие.

Diagnostics, Год журнала: 2024, Номер 14(6), С. 591 - 591

Опубликована: Март 11, 2024

While the adoption of wireless capsule endoscopy (WCE) has been steadily increasing, its primary application remains limited to observing small intestine, with relatively less in upper gastrointestinal tract. However, there is a growing anticipation that advancements technology will lead significant increase examinations. This study addresses underexplored domain landmark identification within tract using WCE, acknowledging research and public datasets available this emerging field. To contribute future development WCE for gastroscopy, novel approach proposed. Utilizing color transfer techniques, simulated dataset tailored created. Using Euclidean distance measurements, similarity between color-transferred authentic images verified. Pioneering exploration anatomical classification data, integrates evaluation image preprocessing deep learning specifically employing DenseNet169 model. As result, utilizing achieves an accuracy exceeding 90% Furthermore, sharpen detail filters demonstrates from 91.32% 94.06%.

Язык: Английский

Процитировано

2
CLASSIFICATION OF ENDOSCOPIC IMAGES USING CNN ARCHITECTURE BASED ON FEATURE INTEGRATION DOI Open Access
Hüseyin Üzen, Hüseyin Fırat

Kahramanmaraş Sütçü İmam Üniversitesi Mühendislik Bilimleri Dergisi, Год журнала: 2024, Номер 27(1), С. 121 - 132

Опубликована: Март 3, 2024

Derin öğrenme (DL) tekniklerindeki son gelişmeler, tıbbi görüntüler kullanılarak gastrointestinal (GI) hastalıkların sınıflandırılmasını otomatikleştirmek için umut verici bir potansiyel göstermektedir. Zamanında ve kesin teşhis, tedavi etkinliğini önemli ölçüde etkilemektedir. Bu araştırma, GI hastalıklarını tanımlamak yeni DL tabanlı modeli tanıtmaktadır. model, önceden eğitilmiş ağ mimarilerinin ara katmanlarından elde edilen öznitelikleri birleştirerek sınıflandırma işlemini gerçekleştirmektedir. Öznitelik entegrasyonuna dayalı evrişimsel sinir ağı (ESA) olarak adlandırılan bu modelde, endoskopik görüntüleri sınıflandırmak yüksek düşük seviyeli birleştirilerek nihai öznitelik haritası edilmektedir. Daha sonra kullanılmaktadır. Kvasirv2 veri seti yapılan deneysel analizler sonucunda, önerilen model ile başarılı performans edilmiştir. Özellikle, DenseNet201 modelinin katmanlarındaki özelliklerin birleştirilmesi, sırasıyla %94.25, %94.28, %94.24 doğruluk, kesinlik, duyarlılık F1 puanı sonuçlanmıştır. Diğer ESA modellerle çalışmalarla karşılaştırmalı analizler, modelin üstünlüğünü ortaya koymuş doğruluğu %94.25'e yükseltmiştir. Bu, görüntülerden hastalık tespitinde gelişmiş DenseNet201'in özelliklerden yararlanma potansiyelinin altını çizmektedir.

Процитировано

1