Skip to main content

Table 8 Comparison of results from the 3 different LipAuth model weights presented in this section, against lighting conditions for the 20 unseen individuals from FAVLIPS

From: Understanding visual lip-based biometric authentication for mobile devices

 XM2VTS onlyUpdated weightsRetrained
XM2VTS: Evaluation Set1.21%1.95%5.60%
FAVLIPS: Neutral Nums22.43%13.79%10.83%
FAVLIPS: Light Front28.44%17.50%20.54%
FAVLIPS: Light Side42.29%36.67%30.00%
FAVLIPS: Light Behind44.91%24.17%29.12%