Sensor-to-Trace : Self-Supervised Learning forMulti-Domain Handwriting Reconstruction fromStabilo DigiPen’s Sensor Signals

we will explore two main avenues for improving existing systems. First, building on the preliminary study, we will investigate the benefit of self-supervised learning in a multi-domain learning framework. In a second step, we will explore new ways of assessing the quality of the reconstruction. More details in the pdf file below !