GestDoor: Gesture-Based User Authentication for Door Entries Utilizing Wearable IMUs

This work introduces GestDoor, a novel behavioral biometric system for door access control, leveraging arm movements during door-opening actions. We compiled an extensive dataset of 3,330 samples—surpassing those in current state-of-the-art studies—with 11 participants wearing two 6-degree-of-freedom (DOF) inertial motion units (IMUs) placed on the wrist and upper arm. Each participant completed up to three data collection sessions, performing four distinct door-opening activities: left-hand pull, left-hand push, right-hand pull, and right-hand push. We extracted various temporal and frequency-domain features to assess authentication performance within sessions and permanence (performance over time). Our evaluation, using three classifiers and dynamic time warping for signal matching, shows that GestDoor achieves an impressive equal error rate (EER) range of 0.0%-0.7%, validating its potential viability as a biometric approach.