The Technology Research Of 3D Face Modeling Based On The Activr Apparent Model

Along with the development of multimedia technologies, the 3D face modeling technology becomes research hot spot in computer vision and so on in the last few years. The accuracy of modeling will immediately influence the following image processing effect. The speed of active apparent model is quick and the model is easy to adjust, therefore this article which is based on the active apparent model algorithm to carry on the analysis, the research and the improvement to the three dimensional face modeling technology has important meanings. This article mainly aiming in the question of the speed and the precision of face modeling, has launched the following research:First aiming in the active apparent model building process the texture vector and the shape vector in the fitting process dimension possibly inconsistent question, the article carry on normalized processing. Then the article through introducing the parameter c realizes the shape vector and the texture vector's organic synthesis and cause good effect. And through adjusting the parameter c, may realize the active apparent model which we establish to be more stable, to be more effective, to be even more satisfying our requirement.Second we combining the Loop segmentation algorithm and the Delaunay triangulation algorithm, carry on the segmentation to the initial productive face shape and the texture grid model, and carry on the optimization the person face grid model using up the Light along technology. We have avoided the morbid state triangle grid's appearance to the greatest degree. From the experimental result we obtain the high third dimension three dimensional person face shape and the texture grid model.Third aiming at the problem of the present active apparent model which has numerous parameters and the objective function is complex and the match speed is slow, this article proposed an improvement genetic algorithm which is based on the direction from the study. The experimental result indicated based on the improvement direction from the study genetic algorithm's prototype matching algorithm. Prototype matching algorithm's robustness and the stability has have enhanced. The algorithm's convergence rate has also been sped up. The model match precision has also been enhanced. The prototype matching time has also been reduced.