DL-SAR-Hotspots is a deep learning framework for predicting Specific Absorption Rate (SAR) hotspot locations and values from electromagnetic simulation data. The project uses Convolutional Neural Networks (CNNs) to (1) localize SAR hotspots within body cross-section images and (2) estimate the corresponding SAR magnitude.
machine-learning medical-imaging convolutional-neural-networks computational-electromagnetics bioelectromagnetics electromagnetic-safety
-
Updated
Feb 27, 2026 - Jupyter Notebook