基于支持向量机的图像深度提取方法

Depth extraction method of images based on support vector machine

在目前的2D-3D视频转换技术的研究中,大多数方案都是利用一种线索或者几种线索简单组合的方法来提取图像的深度信息,导致方案的受场景限制而准确性不高.为了提高深度提取方案的准确性,提出基于支持向量机的图像深度提取方法.该算法通过支持向量机的学习获得各线索合理的结合原则,利用图像中提取的纹理变化、纹理梯度以及雾度等线索进行深度预测,从而提高深度提取算法的准确性,实验结果表明该算法是有效的.

In current 2D-3D video conversion technology research,most programs extract depth information of image by using a single clue or the simple combination of several clues,thus leading the inaccuracy due to the limitation of certain conditions.In order to improve the accuracy of the depth extraction solution,this paper put forward the depth extraction method of images based on support vector machine（SVM）.By the study of support vector machine in order to get the reasonable combination of all the clues,used the texture gradient and fog,and texture changes which were exacted from the images and making deep prediction,the accuracy of the depth extraction algorithm could be improved.The experimental results showed that the algorithm was effective.