Designing a survey to examine an anonymisation method for driver videos

Abstract: Data collected during naturalistic driving studies often includes video of thedriver’s face. Due to the Data Privacy Act (“Personuppgiftslagen” (PuL) in Swedish),it is desirable to find a good method that can make the driver’s face anonymous, whilekeeping the original driver‘s facial expressions. In the unEye project, an attempt tocreate such method was made. Using this method, the original driver’s facial expressionsare coded into action units which are then used to construct a correspondingvideo in which the driver face is replaced by an animated face model.This thesis project was conducted to validate whether the anonymisation methodbuilt in the unEye project could really achieve its objectives. In particular, thisthesis contributes in validating whether a human viewer can identify who the originaldriver was from the animated version of the video and examining how well the facialexpressions are translated when creating the animated driver videos. For this, acomputer-based survey was designed and performed. The survey has two differentparts corresponding to the two different validation tasks.The results from the survey showed some unexpectedly high number of correctclassifications in the identification part. However, some explanations were foundwhich suggest that these high numbers were most likely due to other reasons thanthe ability of the test participants to identify the driver. For the expressions part, theresult showed that surprise was the best preserved expression and that anger was themost poorly preserved one. The average recognition proportion of the expressionswas 24% and some patterns were found in the results which indicate that the detailsin the face are not translated well enough, especially in the regions around the mouth,eyes and eyebrows. In conclusion, while the method is able to make the driver faceanonymous, it should be improved so that the facial expressions are better translated.The results were also examined with respect to the age, gender, area of occupationand education level of the test participants and no clear difference could be foundwithin these groups. This indicates that the result could be generalised to a largerpopulation.