Search papers, labs, and topics across Lattice.
This study introduces an attention-guided deep learning framework that enhances sperm morphology classification by integrating a pretrained EfficientNet-B0 with a Convolutional Block Attention Module (CBAM). The model not only achieves high accuracy—90.2% and 93.9% on the SMIDS and HuSHem datasets, respectively—but also provides interpretability through Grad-CAM++ visualizations, allowing clinicians to understand the decision-making process. By addressing the dual challenges of accuracy and interpretability, this framework presents a viable solution for automated sperm analysis in fertility clinics, potentially improving male infertility diagnosis.
Attention-guided deep learning can achieve over 90% accuracy in sperm morphology classification while providing critical interpretability for clinical applications.
Male infertility is a major cause of couple infertility, often linked to abnormal sperm morphology. While deep learning models offer automated analysis, most lack interpretability, limiting their clinical adoption. This study proposes an attention-guided deep learning framework for sperm morphology classification. We combine a pretrained EfficientNet-B0 with a Convolutional Block Attention Module (CBAM) to focus on key areas of the sperm head, improving both accuracy and interpretability. Evaluated on the SMIDS and HuSHem public datasets, our model achieves accuracies of 90.2% and 93.9% (macro F1 scores of 0.913 and 0.948), outperforming SimpleCNN and standard EfficientNet-B0. Furthermore, we use Grad-CAM++ visualizations to highlight features influencing the model's decisions. The results demonstrate that this accurate and transparent framework is a practical tool for automated sperm analysis in fertility clinics.
