Real Time Smile Detection using Haar Classifiers on SoC

摘要

Smile detection in real time video is an interesting problem with many potential applications. This paper is intended to implement a real time smile detection from video using Haar Classifiers through Raspberry Pi BCM2835 processor which is a combination of SoC with GPU based Architecture. For capturing video we used Raspberry Pi Camera Board of 5MP and which plugs directly into the CSI connector on the Raspberry Pi. Computer vision OpenCV libraries with python IDE is used for face detection and smile detection through linux based raspbian operating system. Frame rates of various video resolutions during smile detection in Raspberry Pi are also observed. The present method can be used in low cost robotic computer vision applications, human computer interaction (HCI) and in education as well because cheap and small hardware used.