Enhancing brain tumor detection in MRI images through explainable AI using Grad-CAM with Resnet 50

BMC Medical Imaging, Journal Year: 2024, Volume and Issue: 24(1)

Published: May 11, 2024

Abstract This study addresses the critical challenge of detecting brain tumors using MRI images, a pivotal task in medical diagnostics that demands high accuracy and interpretability. While deep learning has shown remarkable success image analysis, there remains substantial need for models are not only accurate but also interpretable to healthcare professionals. The existing methodologies, predominantly learning-based, often act as black boxes, providing little insight into their decision-making process. research introduces an integrated approach ResNet50, model, combined with Gradient-weighted Class Activation Mapping (Grad-CAM) offer transparent explainable framework tumor detection. We employed dataset enhanced through data augmentation, train validate our model. results demonstrate significant improvement model performance, testing 98.52% precision-recall metrics exceeding 98%, showcasing model’s effectiveness distinguishing presence. application Grad-CAM provides insightful visual explanations, illustrating focus areas making predictions. fusion explainability holds profound implications diagnostics, offering pathway towards more reliable detection tools.

Language: Английский

---

Integrated approach of federated learning with transfer learning for classification and diagnosis of brain tumor

BMC Medical Imaging, Journal Year: 2024, Volume and Issue: 24(1)

Published: May 15, 2024

Abstract Brain tumor classification using MRI images is a crucial yet challenging task in medical imaging. Accurate diagnosis vital for effective treatment planning but often hindered by the complex nature of morphology and variations Traditional methodologies primarily rely on manual interpretation images, supplemented conventional machine learning techniques. These approaches lack robustness scalability needed precise automated classification. The major limitations include high degree intervention, potential human error, limited ability to handle large datasets, generalizability diverse types imaging conditions.To address these challenges, we propose federated learning-based deep model that leverages power Convolutional Neural Networks (CNN) accurate brain This innovative approach not only emphasizes use modified VGG16 architecture optimized also highlights significance transfer domain. Federated enables decentralized training across multiple clients without compromising data privacy, addressing critical need confidentiality handling. benefits from technique utilizing pre-trained CNN, which significantly enhances its classify tumors accurately leveraging knowledge gained vast datasets.Our trained dataset combining figshare, SARTAJ, Br35H employing decentralized, privacy-preserving training. adoption further bolsters model’s performance, making it adept at handling intricate associated with different tumors. demonstrates precision (0.99 glioma, 0.95 meningioma, 1.00 no tumor, 0.98 pituitary), recall, F1-scores classification, outperforming existing methods. overall accuracy stands 98%, showcasing efficacy classifying various accurately, thus highlighting transformative enhancing images.

Language: Английский

---

Enhancing accessibility for improved diagnosis with modified EfficientNetV2-S and cyclic learning rate strategy in women with disabilities and breast cancer

Frontiers in Medicine, Journal Year: 2024, Volume and Issue: 11

Published: March 7, 2024

Breast cancer, a prevalent cancer among women worldwide, necessitates precise and prompt detection for successful treatment. While conventional histopathological examination is the benchmark, it lengthy process prone to variations different observers. Employing machine learning automate diagnosis of breast presents viable option, striving improve both precision speed. Previous studies have primarily focused on applying various deep models classification images. These methodologies leverage convolutional neural networks (CNNs) other advanced algorithms differentiate between benign malignant tumors from Current models, despite their potential, encounter obstacles related generalizability, computational performance, managing datasets with imbalances. Additionally, significant number these do not possess requisite transparency interpretability, which are vital medical diagnostic purposes. To address limitations, our study introduces an model based EfficientNetV2. This incorporates state-of-the-art techniques in image processing network architecture, aiming accuracy, efficiency, robustness classification. We employed EfficientNetV2 model, fine-tuned specific task Our underwent rigorous training validation using BreakHis dataset, includes diverse Advanced data preprocessing, augmentation techniques, cyclical rate strategy were implemented enhance performance. The introduced exhibited remarkable efficacy, attaining accuracy 99.68%, balanced recall as indicated by F1 score, considerable Cohen’s Kappa value. indicators highlight model’s proficiency correctly categorizing images, surpassing current reliability effectiveness. research emphasizes improved accessibility, catering individuals disabilities elderly. By enhancing visual representation proposed approach aims make strides inclusive interpretation, ensuring equitable access information.

Language: Английский

---

Optimizing double-layered convolutional neural networks for efficient lung cancer classification through hyperparameter optimization and advanced image pre-processing techniques

BMC Medical Informatics and Decision Making, Journal Year: 2024, Volume and Issue: 24(1)

Published: May 27, 2024

Abstract Lung cancer remains a leading cause of cancer-related mortality globally, with prognosis significantly dependent on early-stage detection. Traditional diagnostic methods, though effective, often face challenges regarding accuracy, early detection, and scalability, being invasive, time-consuming, prone to ambiguous interpretations. This study proposes an advanced machine learning model designed enhance lung stage classification using CT scan images, aiming overcome these limitations by offering faster, non-invasive, reliable tool. Utilizing the IQ-OTHNCCD dataset, comprising scans from various stages healthy individuals, we performed extensive preprocessing including resizing, normalization, Gaussian blurring. A Convolutional Neural Network (CNN) was then trained this preprocessed data, class imbalance addressed Synthetic Minority Over-sampling Technique (SMOTE). The model’s performance evaluated through metrics such as precision, recall, F1-score, ROC curve analysis. results demonstrated accuracy 99.64%, F1-score values exceeding 98% across all categories. SMOTE enhanced ability classify underrepresented classes, contributing robustness These findings underscore potential in transforming diagnostics, providing high classification, which could facilitate detection tailored treatment strategies, ultimately improving patient outcomes.

Language: Английский

---

Advanced AI-driven approach for enhanced brain tumor detection from MRI images utilizing EfficientNetB2 with equalization and homomorphic filtering

BMC Medical Informatics and Decision Making, Journal Year: 2024, Volume and Issue: 24(1)

Published: April 30, 2024

Brain tumors pose a significant medical challenge necessitating precise detection and diagnosis, especially in Magnetic resonance imaging(MRI). Current methodologies reliant on traditional image processing conventional machine learning encounter hurdles accurately discerning tumor regions within intricate MRI scans, often susceptible to noise varying quality. The advent of artificial intelligence (AI) has revolutionized various aspects healthcare, providing innovative solutions for diagnostics treatment strategies. This paper introduces novel AI-driven methodology brain from images, leveraging the EfficientNetB2 deep architecture. Our approach incorporates advanced preprocessing techniques, including cropping, equalization, application homomorphic filters, enhance quality data more accurate detection. proposed model exhibits substantial performance enhancement by demonstrating validation accuracies 99.83%, 99.75%, 99.2% BD-BrainTumor, Brain-tumor-detection, Brain-MRI-images-for-brain-tumor-detection datasets respectively, this research holds promise refined clinical patient care, fostering reliable identification images. All is available Github: https://github.com/muskan258/Brain-Tumor-Detection-from-MRI-Images-Utilizing-EfficientNetB2 ).

Language: Английский

---

Enhanced skin cancer diagnosis using optimized CNN architecture and checkpoints for automated dermatological lesion classification

BMC Medical Imaging, Journal Year: 2024, Volume and Issue: 24(1)

Published: Aug. 2, 2024

Abstract Skin cancer stands as one of the foremost challenges in oncology, with its early detection being crucial for successful treatment outcomes. Traditional diagnostic methods depend on dermatologist expertise, creating a need more reliable, automated tools. This study explores deep learning, particularly Convolutional Neural Networks (CNNs), to enhance accuracy and efficiency skin diagnosis. Leveraging HAM10000 dataset, comprehensive collection dermatoscopic images encompassing diverse range lesions, this introduces sophisticated CNN model tailored nuanced task lesion classification. The model’s architecture is intricately designed multiple convolutional, pooling, dense layers, aimed at capturing complex visual features lesions. To address challenge class imbalance within an innovative data augmentation strategy employed, ensuring balanced representation each category during training. Furthermore, optimized layer configuration augmentation, significantly boosting precision detection. learning process using Adam optimizer, parameters fine-tuned over 50 epochs batch size 128 ability discern subtle patterns image data. A Model Checkpoint callback ensures preservation best iteration future use. proposed demonstrates 97.78% notable 97.9%, recall F2 score 97.8%, underscoring potential robust tool classification cancer, thereby supporting clinical decision-making contributing improved patient outcomes dermatology.

Language: Английский

---

Artificial Intelligence in Head and Neck Cancer: Innovations, Applications, and Future Directions

Current Oncology, Journal Year: 2024, Volume and Issue: 31(9), P. 5255 - 5290

Published: Sept. 6, 2024

Artificial intelligence (AI) is revolutionizing head and neck cancer (HNC) care by providing innovative tools that enhance diagnostic accuracy personalize treatment strategies. This review highlights the advancements in AI technologies, including deep learning natural language processing, their applications HNC. The integration of with imaging techniques, genomics, electronic health records explored, emphasizing its role early detection, biomarker discovery, planning. Despite noticeable progress, challenges such as data quality, algorithmic bias, need for interdisciplinary collaboration remain. Emerging innovations like explainable AI, AI-powered robotics, real-time monitoring systems are poised to further advance field. Addressing these fostering among experts, clinicians, researchers crucial developing equitable effective applications. future HNC holds significant promise, offering potential breakthroughs diagnostics, personalized therapies, improved patient outcomes.

Language: Английский

---

Enhancing brain tumor classification in MRI scans with a multi-layer customized convolutional neural network approach

Frontiers in Computational Neuroscience, Journal Year: 2024, Volume and Issue: 18

Published: June 12, 2024

The necessity of prompt and accurate brain tumor diagnosis is unquestionable for optimizing treatment strategies patient prognoses. Traditional reliance on Magnetic Resonance Imaging (MRI) analysis, contingent upon expert interpretation, grapples with challenges such as time-intensive processes susceptibility to human error.

Language: Английский

---

Revolutionizing breast ultrasound diagnostics with EfficientNet-B7 and Explainable AI

BMC Medical Imaging, Journal Year: 2024, Volume and Issue: 24(1)

Published: Sept. 2, 2024

Breast cancer is a leading cause of mortality among women globally, necessitating precise classification breast ultrasound images for early diagnosis and treatment. Traditional methods using CNN architectures such as VGG, ResNet, DenseNet, though somewhat effective, often struggle with class imbalances subtle texture variations, to reduced accuracy minority classes malignant tumors. To address these issues, we propose methodology that leverages EfficientNet-B7, scalable architecture, combined advanced data augmentation techniques enhance representation improve model robustness. Our approach involves fine-tuning EfficientNet-B7 on the BUSI dataset, implementing RandomHorizontalFlip, RandomRotation, ColorJitter balance dataset The training process includes stopping prevent overfitting optimize performance metrics. Additionally, integrate Explainable AI (XAI) techniques, Grad-CAM, interpretability transparency model's predictions, providing visual quantitative insights into features regions influencing outcomes. achieves 99.14%, significantly outperforming existing CNN-based approaches in image classification. incorporation XAI enhances our understanding decision-making process, thereby increasing its reliability facilitating clinical adoption. This comprehensive framework offers robust interpretable tool detection cancer, advancing capabilities automated diagnostic systems supporting processes.

Language: Английский

---

Vision Transformers for Low-Quality Histopathological Images: A Case Study on Squamous Cell Carcinoma Margin Classification

Diagnostics, Journal Year: 2025, Volume and Issue: 15(3), P. 260 - 260

Published: Jan. 23, 2025

Background/Objectives: Squamous cell carcinoma (SCC), a prevalent form of skin cancer, presents diagnostic challenges, particularly in resource-limited settings with low-quality imaging infrastructure. The accurate classification SCC margins is essential to guide effective surgical interventions and reduce recurrence rates. This study proposes vision transformer (ViT)-based model improve margin by addressing the limitations convolutional neural networks (CNNs) analyzing histopathological images. Methods: introduced transfer learning approach using ViT architecture customized additional flattening, batch normalization, dense layers enhance its capability for classification. A performance evaluation was conducted machine metrics averaged over five-fold cross-validation comparisons were made leading CNN models. Ablation studies have explored effects architectural configuration on performance. Results: ViT-based achieved superior 0.928 ± 0.027 accuracy 0.927 0.028 AUC, surpassing highest performing model, InceptionV3 (accuracy: 0.86 0.049; AUC: 0.837 0.029), demonstrating robustness reinforced importance tailored configurations enhancing Conclusions: underscores transformative potential ViTs analysis, especially settings. By reducing dependence high-quality specialized expertise, it scalable solution global cancer diagnostics. Future research should prioritize optimizing such environments broadening their clinical applications.

Language: Английский

---