Finger Contact Forces from Thermal Signatures

Contact forces are essential for understanding manipulation, tactile interaction, and robotic control. Measuring these forces, however, remains challenging due to the limitations of conventional sensors, which are often costly, geometry-dependent, and impractical for complex interactions. Existing datasets such as ContactDB and GRAB provide rich 3D and contact-area information but lack explicit quantitative force data. In this project, we present a novel approach to infer contact forces from thermal imaging, bridging the gap between visible thermal signatures and physical interaction strength. We present a dataset comprising thermal contact patterns of finger pads recorded over multiple time periods, including precise force measurements up to a range of \SI{20}{N}. Furthermore, we evaluate the relationship between thermal signatures and single-finger forces, establishing a foundation for data-driven models toward non-intrusive force estimation in human–object interaction research.