An efficient smile and frown detection algorithm

This paper proposed a multi-people simile and frown detection with a novel method to detect frown expression. The proposed grayscale analysis is more suitable for implementing in handheld devices which are capable torecognize multi-people in real time. The OPENCV is applied in the proposed method for face detection, and whole system is realized on both computer and android devices, respectively. Comparing with the previous SOBEL algorithm, the processing speed is greatly improved that also remains the recognition accuracy. The database design is downloaded from Facebook to fit the real application. In theexperimentof ten people real-timetest, the smile and frown recognition can reach 80% and 90%, respectively. And for the database test, the smile and frown recognitionof multi-people can reach 72.7% and 78.8%, respectively either.