由于人手形状复杂、细节丰富，构建个性化的人手模型充满挑战。提出一种基于立体视觉和RBF几何变形技术的个性化人手建模方法，通过综合采用轮廓点和SIFT点描述手部特征，提取用户手部3维特征点云，并以点云为约束，采用几何对齐两步法对齐手模板，结合RBF插值算法，得到具有用户特征的个性化手模型。实验结果表明，该方法可以有效地重建出用户手的几何模型。

It is very challenging to construct a personalized virtual hand because of the complex shape and rich details of a human hand. An approach is proposed based on stereo vision and RBF interpolation to construct a personalized virtual hand model. Contour points and SIFT points are integrated to depict the features of a human hand, while a three-dimensional point cloud is extracted via stereo vision. A two-pass geometry alignment is performed to match a three-dimensional hand template to the point features. Guided by the point features, RBF interpolation is employed to deform the hand template to produce the personalized virtual hand model. Experimental results show that the proposed approach is functional enough to reconstruct the geometry of user’s hand.