Abstract: This paper proposes a feature extraction algorithm based on elevation. A 3D model is represented with six gray-level images which describes the altitude information of 2D projections from six different views including front, left, right, rear, top and bottom. Each gray-level image, called an elevation, is decomposed into a set of concentric circles. The sum of the altitude information within each concentric circle is calculated. The elevation descriptor is obtained from the difference of the altitude sums between two successive concentric circles. Experimental results demonstrate the retrieval effectiveness of the method is considerable with LFD method, and it calculates faster.

Key words: 3D model retrieval, elevation, feature extraction

摘要： 提出一种基于正视图的三维模型特征提取算法，从前、顶、右、后、底与左分别对模型进行投影，获得6个正视图，使用不同半径的同心圆将每个正视图分割，计算每个同心圆内的高度信息和，得到相邻2个同心圆内高度信息和的差值，每个三维模型特征由 6个正视图描述。实验结果证明，该方法的检索效果与光场法相当，计算速度较快。

关键词: 三维模型检索, 正视图, 特征提取

CLC Number: 

• TP391.4