Development And Analysis Of Face Recognition System On A Mobile Robot Environment

Quah, Jit Shen (2016) Development And Analysis Of Face Recognition System On A Mobile Robot Environment. Project Report. UTeM, Melaka, Malaysia . (Submitted)

Text (24 Pages) Development And Analysis Of Face Recognition System On A Mobile Robot Environment.pdf - Submitted Version Download (156kB)

Abstract

Surveillance technologies has become more and more common especially in household due to its versatility and applicability however traditional surveillance technology such as the CCTV camera suffers from its traditional problems such as having blinds spots around the area outside of the camera. Many approaches has already been made to solve said issue mainly by giving multiple degree of freedom but it is clear that some levels of said problems still maintain thus our approached in solving said issue is mobilizing said camera itself and giving it face recognition abilities. However, the main problem with this idea is that face recognition algorithm although extensively developed is done almost exclusively on a static base thus its performance in terms of accuracy and speed of detection on a mobile environment is mostly unknown. The main difference between static and stationary is the effect of blur which is a form of noise usually ignored due to the stationary state of the camera. Said blur can cause issues especially when it comes to doing detailed feature analysis such as face recognition because the blur would increase the size of the boundary line causing camera to may or may not detect the face at all or detect it as something entirely alien. Thus from this it is safe to hypothesize that the performances of said face recognition algorithm may it be eigenvalues or fisherface weight map’s performance would differ negatively when comparing its results of a static base to a mobile base and to prove it, we have run performance analysis in terms of range of capture for face identification and recognition by moving the robot while trying to identify faces at different duty cycles. The results from this research shows that when there is an increase in mobile robot speed from 0 ~ 65% duty cycle there seem to be a reduction in performance in terms of range of capture of approximately 30% for both face recognition and face identification which is a clear reduction in performance.

Actions (login required)

View Item