Identifying Induced Polarization Effects in TEM Data by Deep Learning

Summary This study presents a deep learning framework for identifying induced polarization (IP)-affected transient electromagnetic (TEM) data. The three-stage framework categorizes data into those that can be modeled under simpler resistivity-only conditions, those requiring the inclusion of IP parameters, and those unsuitable for either approach. The framework was tested on a tTEM dataset from Kigadye, Tanzania, where the results demonstrated strong alignment with manual assessments, showcasing its potential for automating TEM data analysis.