AINeutralarXiv – CS AI · 6h ago6/10
🧠
Interpretable Sperm Morphology Classification via Attention-Guided Deep Learning
Researchers developed an interpretable deep learning framework using EfficientNet-B0 and attention mechanisms to classify sperm morphology for male infertility diagnosis. The model achieves 90-94% accuracy on public datasets while providing visual explanations through Grad-CAM++ visualizations, addressing the clinical adoption barrier of traditional black-box AI models.