LBNet: an optimized lightweight CNN for mammographic breast cancer classification with XAI-based interpretability

Keyphrases

• Cancer Cases 100%
• Mammogram Images 100%
• Breast Cancer Classification 100%
• Lightweight Neural Network 100%
• Breast Cancer 50%
• Socioeconomic 50%
• Health Burden 50%
• Deep Learning 50%
• Incidence Rate 50%
• Feature Extraction 50%
• Strong Potential 50%
• High Incidence 50%
• Max Pooling 50%
• Computational Efficiency 50%
• Mortality Rate 50%
• Computational Complexity 50%
• Convolutional Layer 50%
• External Data 50%
• Global Health 50%
• Shapley Additive Explanations 50%
• 5-fold Cross Validation 50%
• Non-cancer 50%
• Breast Cancer Detection 50%
• VGG19 50%
• High Mortality 50%
• Training Parameters 50%
• Mammogram 50%
• Batch Normalization 50%
• Relevant Region 50%
• Resource-constrained Settings 50%
• Adam Optimizer 50%
• Interpretable Convolutional Neural Network 50%
• CBIS-DDSM 50%
• Mammography 50%
• Breast Cancer Screening 50%
• ResNet152V2 50%
• Gradient-weighted Class Activation Mapping (Grad-CAM) 50%
• Improved Prediction 50%
• Transfer Learning Models 50%
• Low Accuracy 50%
• SE-ResNet152 50%
• Clinical Trust 50%
• Clinical Deployment 50%
• ReLU Activation Function 50%

Computer Science

• Interpretability 100%
• Explainable Artificial Intelligence 100%
• Convolutional Neural Network 100%
• Generalizability 100%
• Learning Approach 50%
• Fold Cross Validation 50%
• Computational Complexity 50%
• Transfer Learning 50%
• Convolutional Layer 50%
• Computational Efficiency 50%
• Baseline Model 50%
• Gradient-Weighted Class Activation Mapping 50%
• Batch Normalization 50%
• Feature Extraction 50%
• Deep Learning Method 50%
• VGG-19 Convolutional Neural Network 50%
• SHapley Additive exPlanation 50%
• Adam Optimizer 50%
• Cancer Diagnosis 50%

Biochemistry, Genetics and Molecular Biology

• Mortality Rate 100%
• Cancer Classification 100%
• Feature Extraction 100%
• Breast Cancer Screening 100%