欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
视差图的获取既是立体图像处理的核心，也是立体图像处理的难点。为了高效、精确地获取立体图像的视差图，提出了一种基于Panum融合区的视差图获取算法。首先根据人眼视觉特性的Panum融合区得到立体图像视差敏感范围的计算方法；然后根据此范围将立体图像分割成视差敏感区和视差不敏感区，对视差不敏感区计算低精度视差图，对视差敏感区计算高精度视差图；最后将两视差图相结合得到最终的视差图。实验结果表明，本文所提出的算法在不影响人眼主观立体感受的情况下，计算出较高精度的视差图，同时相对于通常的高精度视差图获取算法降低了时间开销。
英文摘要：
Extracting disparity map as a difficult problem is the core of stereo image processing. A disparity map calculating algorithm based on Panum＇s fusional area is proposed for extracting disparity map efficiently and accurately. In this paper, we first derive the calculation method of susceptible disparity range of stereo images according to Panum＇s fusional area-an important characteristic area in human vision. And then based on the susceptible disparity range,the stereo image is segmented into insusceptible dis- parity zone and susceptible disparity zone. Next, the low-accuracy disparity map for insusceptible disparity zone and the high-accuracy disparity map for susceptible disparity zone are calculated. At last, two types of accurate disparity maps are combined into the last disparity map. The experimental results show that the proposed algorithm in this paper can extract high-accuracy disparity map without influencing the subiective stereo experience of human vision,and can reduce the time cost compared with ordinary algorithm for calculating high-accuracy disparity map.
同期刊论文项目
同项目期刊论文
Rate-distortion optimized mode switching for error-resilient multi-view video plus depth based 3-D v
Rate distortion optimized mode switching for error-resilient multi-view video plus depth based 3D vi
A stereo image quality evaluation method based on three dimensional characteristics and structural s
期刊信息
位置：