An amalgamation of deep neural networks optimized with Salp swarm algorithm for cervical cancer detection

Computers & Electrical Engineering, Год журнала: 2025, Номер 123, С. 110106 - 110106

Опубликована: Янв. 28, 2025

Язык: Английский

---

A hybrid framework for colorectal cancer detection and U-Net segmentation using polynetDWTCADx

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

Опубликована: Янв. 5, 2025

"PolynetDWTCADx" is a sophisticated hybrid model that was developed to identify and distinguish colorectal cancer. In this study, the CKHK-22 dataset, comprising 24 classes, served as introduction. The proposed method, which combines CNNs, DWTs, SVMs, enhances accuracy of feature extraction classification. study employs DWT optimize enhance two integrated CNN models before classifying them with SVM following systematic procedure. PolynetDWTCADx most effective we evaluated. It capable attaining moderate level recall, well an area under curve (AUC) during testing. testing 92.3%, training 95.0%. This demonstrates distinguishing between noncancerous cancerous lesions in colon. We can also employ semantic segmentation algorithms U-Net architecture accurately segment regions. assessed model's exceptional success segmenting providing precise delineation malignant tissues using its maximal IoU value 0.93, based on intersection over union (IoU) scores. When these techniques are added PolynetDWTCADx, they give doctors detailed visual information needed for diagnosis planning treatment. These very good at finding separating has potential recognition management cancer, underscores.

Язык: Английский

---

Colorectal cancer detection with enhanced precision using a hybrid supervised and unsupervised learning approach

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

Опубликована: Янв. 25, 2025

The current work introduces the hybrid ensemble framework for detection and segmentation of colorectal cancer. This will incorporate both supervised classification unsupervised clustering methods to present more understandable accurate diagnostic results. method entails several steps with CNN models: ADa-22 AD-22, transformer networks, an SVM classifier, all inbuilt. CVC ClinicDB dataset supports this process, containing 1650 colonoscopy images classified as polyps or non-polyps. best performance in ensembles was done by AD-22 + Transformer model, AUC 0.99, a training accuracy 99.50%, testing 99.00%. group also saw high 97.50% Polyps 99.30% Non-Polyps, together recall 97.80% 98.90% hence performing very well identifying cancerous healthy regions. proposed here uses K-means combination visualisation bounding boxes, thereby improving yielding silhouette score 0.73 cluster configuration. It discusses how combine feature interpretation challenges into medical imaging localization precise malignant A good balance between generalization shall be hyperparameter optimization-heavy learning rates; dropout rates overfitting suppressed effectively. schema treats deficiencies previous approaches, such incorporating CNN-based effective extraction, networks developing attention mechanisms, finally fine decision boundary support vector machine. Further, we refine process via purpose enhancing procedure. Such holistic framework, hence, further boosts results generating outcomes rigorous benchmarking detecting cancer higher reality towards clinical application feasibility.

Язык: Английский

---

Differential evolution-driven optimized ensemble network for brain tumor detection

International Journal of Machine Learning and Cybernetics, Год журнала: 2025, Номер unknown

Опубликована: Апрель 16, 2025

Язык: Английский

---

DeepGenMon: A Novel Framework for Monkeypox Classification Integrating Lightweight Attention-Based Deep Learning and a Genetic Algorithm

Diagnostics, Год журнала: 2025, Номер 15(2), С. 130 - 130

Опубликована: Янв. 8, 2025

Background: The rapid global spread of the monkeypox virus has led to serious issues for public health professionals. According related studies, and other types skin conditions can through direct contact with infected animals, humans, or contaminated items. This disease cause fever, headaches, muscle aches, enlarged lymph nodes, followed by a rash that develops into lesions. To facilitate early detection monkeypox, researchers have proposed several AI-based techniques accurately classifying identifying condition. However, there is still room improvement detect classify cases. Furthermore, currently pre-trained deep learning models consume extensive resources achieve accurate classification monkeypox. Hence, these often need significant computational power memory. Methods: paper proposes novel lightweight framework called DeepGenMonto various diseases, such as chickenpox, melasma, others. suggested leverages an attention-based convolutional neural network (CNN) genetic algorithm (GA) enhance accuracy while optimizing hyperparameters model. It first applies attention mechanism highlight assign weights specific regions image are relevant model's decision-making process. Next, CNN employed process visual input extract hierarchical features data multiple classes. Finally, CNN's adjusted using robustness accuracy. Compared state-of-the-art (SOTA) models, DeepGenMon design requires significantly lower easier train few parameters. Its effective integration GA further enhances its performance, making it particularly well suited low-resource environments. evaluated on two datasets. dataset comprises 847 images diverse second contains 659 classified categories. Results: model demonstrates superior performance compared SOTA across key evaluation metrics. On 1, achieves precision 0.985, recall 0.984, F-score 0.985. Similarly, 2, attains 0.981, 0.982, 0.982. Moreover, findings demonstrate ability inference time 2.9764 s 1 2.1753 2. Conclusions: These results also show DeepGenMon's effectiveness in different conditions, highlighting potential reliable tool clinical settings.

Язык: Английский

---

Real-Time Object Detector for Medical Diagnostics (RTMDet): A High-Performance Deep Learning Model for Brain Tumor Diagnosis

Bioengineering, Год журнала: 2025, Номер 12(3), С. 274 - 274

Опубликована: Март 11, 2025

Background: Brain tumor diagnosis requires precise and timely detection, which directly impacts treatment decisions patient outcomes. The integration of deep learning technologies in medical diagnostics has improved the accuracy efficiency these processes, yet real-time processing remains a challenge due to computational intensity current models. This study introduces Real-Time Object Detector for Medical Diagnostics (RTMDet), aims address limitations by optimizing convolutional neural network (CNN) architectures enhanced speed accuracy. Methods: RTMDet model incorporates novel depthwise blocks designed reduce load while maintaining diagnostic precision. effectiveness was evaluated through extensive testing against traditional modern CNN using comprehensive imaging datasets, with focus on capabilities. Results: demonstrated superior performance detecting brain tumors, achieving higher compared existing model’s validated its ability process large datasets real time without sacrificing required reliable diagnosis. Conclusions: represents significant advancement application diagnostics. By balance between precision, enhances capabilities imaging, potentially improving outcomes faster more accurate detection. offers promising solution clinical settings where rapid are critical.

Язык: Английский

---

Optimizing Cervical Lesion Detection Using Deep Learning with Particle Swarm Optimization

Опубликована: Фев. 18, 2025

Язык: Английский

---

Improved gated recurrent unit-based osteosarcoma prediction on histology images: a meta-heuristic-oriented optimization concept

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

Опубликована: Апрель 1, 2025

The major prevalent primary bone cancer is osteosarcoma. Preoperative chemotherapy accompanied by resection as part of the normal course treatment. diagnosis and treatment patients are based on reaction. Contrarily, without operation results in persistent an osteosarcoma regrowth. Thus, should receive comprehensive therapy, which includes tumor-free surgery global chemotherapy, to improve their survival. Hence, early individualized care essential since they may lead more effective therapies higher survival rates. Here, main goal recommended research use a unique deep learning approach predict histology images. Initially, data collected from navigation confluence mobile UT Southwestern/UT Dallas dataset. Next, pre-processing images accomplished Weiner filter technique. Further, segmentation for pre-processed done 2D Otsu's method. From segmented images, features extracted linear discriminant analysis (LDA) approach. These undergo final prediction phase that novel improved recurrent gated unit (IGRU), parameter tuning GRU osprey optimization algorithm (OOA) with consideration error minimization objective function. On contrast various conventional methods, simulation findings demonstrate effectiveness developed model terms numerous analysis.

Язык: Английский

---

Multi-Class Brain Malignant Tumor Diagnosis in Magnetic Resonance Imaging Using Convolutional Neural Networks

Brain Research Bulletin, Год журнала: 2025, Номер unknown, С. 111329 - 111329

Опубликована: Апрель 1, 2025

To reduce the clinical misdiagnosis rate of glioblastoma (GBM), primary central nervous system lymphoma (PCNSL), and brain metastases (BM), which are common malignant tumors with similar radiological features, we propose a new CNN-based model, FoTNet. The model integrates frequency-based channel attention layer Focal Loss to address class imbalance issue caused by limited data available for PCNSL. A multi-center MRI dataset was constructed collecting integrating from Zhejiang University School Medicine's Sir Run Shaw Hospital, along public datasets UPENN TCGA. includes T1-weighted contrast-enhanced (T1-CE) images 58 GBM, 82 PCNSL, 269 BM cases, were divided into training testing sets in 5:2 ratio. FoTNet achieved classification accuracy 92.5% an average AUC 0.9754 on test set, significantly outperforming existing machine learning deep methods distinguishing between BM. Through multiple validations, has proven be effective robust tool accurately classifying these tumors, providing strong support preoperative diagnosis assisting clinicians making more informed treatment decisions.

Язык: Английский

---

Ensemble Architecture of Vision Transformer and CNNs for Breast Cancer Tumor Detection From Mammograms

International Journal of Imaging Systems and Technology, Год журнала: 2025, Номер 35(3)

Опубликована: Апрель 18, 2025

ABSTRACT Addressing the complexities of classifying distinct object classes in computer vision presents several challenges, including effectively capturing features such as color, form, and tissue size for each class, correlating class vulnerabilities, singly features, predicting labels accurately. To tackle these issues, we introduce a novel hybrid deep dense learning technique that combines transfer with transformer architecture. Our approach utilizes ResNet50, EfficientNetB1, our proposed ProDense block backbone models. By integrating Vit‐L16 transformer, can focus on relevant mammography extract high‐value pair offering two alternative methods feature extraction. This allows model to adaptively shift region interest towards type slides. The architecture, particularly Vit‐L16, enhances extraction by efficiently long‐range dependencies data, enabling better understand context relationships between features. aids more accurate classification, especially when fine‐tuning pretrained models, it helps adapt specific characteristics target dataset while retaining valuable information learned from pretraining phase. Furthermore, employ stack ensemble leverage both extension training extensive breast cancer classification. process employed refine layers, enhancing classification accuracy. Evaluating method INbreast dataset, observe significant improvement binary category, outperforming current state‐of‐the‐art classifier 98.08% terms

Язык: Английский

---