Mouth Detection Based on Interest Point

PCA had been used on detection of face successfully before. Although PCA had also been used on detection of mouth and eye, it has low detection rate on mouth detection. Because mouth has more pose than eye and PCA can be affected by illumination easily. Interest point has been a popular way on object detection because of its stable property. This paper use Harris corner interest point on mouth detection with PCA probability model. The experiment result shows that the correct rate and the accuracy of mouth position have all been enhanced.