A learning-based THz multi-layer imaging has been recently used for contactless three-dimensional (3D) positioning and encoding. We show a proof-of-concept demonstration of an emerging quantum machine learning (QML) framework to deal with depth variation, shadow effect, and double-sided content recognition, through an experimental validation.