Objectives: The study focuses on the potential use of Explainable AI (XAI) in improving the detection of HIE using ultrasound in neonates born in Jharkhand, India.
Methods: The study involves using appropriate deep Convolutional Neural Network (CNN) Models on ultrasound images. Later, hybrid models with transfer learning approach will be used to ensure better accuracy.
Expected Outcome: This approach builds trust and confidence in AI-based tools used in medical diagnosis and treatment. The use of XAI in improving the detection of HIE in this population can potentially save lives and improve long-term outcomes for neonates with HIE. At the later stage, we will build a user-friendly Computer-Aided interface for the doctors so that they can simply upload the image and get the result. Furthermore, XAI based tool will assist the doctor in distinguishing typical brain variations with HIE findings. Different modes of the proposed user interface are given as follows:
1. Offline: Lightweight IoT and smartphone (with or without internet) deployable XAI models for detection of HIE.
2. Online: A web application deployed on a Cloud server for the detection of HIE.