Student researcher aims to reduce bias in automated surveillance