The image quality of personal face is the fundamental factor that affects face perceptionsystem performance, however the traditional single camera collection technology can’tguarantee the quality of the face image, and then directly reduces the value of faceinformation in the practical application. Therefore, the research on the best face imageacquisition technology not only has important theoretical significance, but also has importantpractical value.This article presents a thorough study of multi-camera cooperative collection and faceimage quality evaluation techniques, and constructs a complete and practical optimum faceacquisition system with multi-camera. Specific content as follows:First, I research the technology of face pose estimation deeply. In order to enhance theaccuracy of pose estimation, I improve the pose estimation algorithm through evaluation otherangle that has corresponding relationship with face rotation.Secondly, I put forward the multi-camera collaborative acquisition algorithm. Inalgorithm, I set multiple cameras to collect the same area synchronously. In order to selectoptimal view, I compare the color skin share of image that each camera gathered at the sametime. All of the best perspective face images need to get through the pose estimation, thenpick out a set of frontal face images.Finally, I in-depth study the technology of objective face image quality evaluation. Inorder to enhance the coherence of subjective and objective evaluation algorithms, andincrease the accuracy of evaluation result for face recognition, I put forward the method ofimproved image quality evaluation algorithm. I increase face similarity as an indicator ofimage quality and adopt a from coarse to fine evaluation framework. |