| With the development and popularization of electronic mobile devices, mobile phone, tablet computer and other electronic mobile devices have a more and more role in people’s life. In order to solves the problems that user’s eyes are too close to the screen a monocular measurement method based pupil location is proposed. Firstly, face images are captured by the camera and eyes are detected via sparse representation. Secondly, hybrid project function is used to narrow the eye area locate the center of the eyes. Lastly, combined with the actual interpupillary distance and the camera focal length with geometric monocular measurement of similarity method, the distance between the human eye and screen distance is calulated. In this paper, the main research work is following:(1) In this paper, a robust eye detection algorithm via sparse representation is proposed, which regards eye detection as a typical two-class classification problem to detects eyes based on the residual error of sparse representation. Sparse representation of an eye originates from the idea that an eye can be regarded as a linear combination of the training eyes. A new eye can be represented as a sparse linear combination of these training eyes.(2)Using the gradient projection function to further reduce the eye region,due to the gray value is smaller than the human eye area skin, gray eyebrow region, the eye regions have greater response on the gradient projection integral. Finally, with the rectangular integral operator of variance traverse the eye area, the rectangle is found with the minimum variance. The center of eyes are determined by the center of rectangle.(3) Introduces and analyzes current strengths and weaknesses of various monocular measurement. The thesis proposed that using the interpupillary distance in the eye image through eye location combined with the camera focal length. The actual interpupillary distance via pinhole model is calculated the distance between eyes and and the camera. |
PDF Full Text Request