Problem Statement:
To develop a non-invasive deep learning-based system for accurately detecting and analyzing anemia in pregnant women using hematological parameters across multiple clinical centers.
Strategy/ Approach/ Technique/ Experiments used to solve
Strategy: Develop a non-invasive diagnostic framework combining deep learning with hematological indicators. Approach: Leverage multicentric clinical datasets and feature engineering for robust anemia detection. Technique: Use deep neural networks trained on CBC parameters and clinical labels for classification. Experiments: Perform training, validation, and testing across geographically diverse data sources to assess model accuracy and generalizability.
Applications:
Early detection of anemia in pregnant women without the need for invasive procedures.
Clinical decision support for obstetricians using AI-assisted diagnosis.
Improved maternal health outcomes through timely intervention based on model predictions.
Integration into mobile health platforms for rural and low-resource settings.
Standardization of anemia screening across multiple healthcare centers using a unified AI model.