An Intelligent DCNN Approach for Fetal mortality rate and mother risk Detection: Integrating Pathological Reports and Clinical History

Abstract

Fetal mortality and maternal risk pose significant challenges in healthcare, particularly in low-resource settings. According to the Indian Council of Medical Research (ICMR), maternal mortality in India remains a critical concern, with 130 deaths per 100,000 live births, while the World Health Organization (WHO) estimates a global maternal mortality rate of 211 per 100,000 live births. Existing models in fetal and maternal risk prediction often provide only partial insights, emphasizing the need for more comprehensive solutions. This paper proposes an intelligent Deep Convolutional Neural Network (DCNN) model that integrates pathological reports and clinical history for accurate detection and classification of fetal mortality risk and maternal complications. The model leverages hierarchical feature learning through convolutional, pooling, and fully connected layers to automatically process complex medical data. Training on a dataset of 15,000 samples with 30 epochs and a batch size of 16 resulted in a mean average precision (mAP) of 0.971. Real-time testing through a cloud-edge application yielded excellent performance metrics, including an accuracy of 97.85%, recall of 97.67%, throughput of 97.91%, and sensitivity of 97.88%. These results represent a significant improvement over existing methods, demonstrating the potential of our DCNN approach to enhance early detection of fetal mortality risks and maternal health complications.