Abstract: The triangle-to-triangle intersection test is a basic component of all collision detection data structures and algorithms.This paper based on the one-class （SVM） Support Vector Machines presents a new method for testing whether two triangles embedded in three dimensions intersect.First，it transforms a triangle named T_a into hypersphere by kernel，core for a radius of R.Then，according to the relationship between the point-to-point distance which is some points in another triangle named T_b to the core and the radius of R，judge whether or not there is some points in the hypersphere.If exists，it can be concluded that an intersection occurs between two triangles，whereas it is not.Theoretical analysis and experimental results show that the algorithm is very fast and more efficient，which can meet the requirement of real-time interactive collision detection.

Key words: triangle and triangle intersection test, collision detection, one-calss （SVM） Support Vector Machines, kernel

摘要： 三角形和三角形相交测试是碰撞检测数据结构和算法的基本组成部分，基于支持向量机的一类分类方法对三维空间中三角形和三角形相交测试提出了一种新的算法，首先用核函数把其中一个三角形（记为T_a）训练成球心为a半径为R的超球体，然后依据另一个三角形（记为T_b上的某些点到超球体的球心a的距离d_i（i=1，2，…，n）与R的关系，判断这些点是否在超球体内。如果T_b有点在超球体内，则断定两个三角形发生相交，反之则没有。理论分析和实验结果都表明，该算法速度很快，效率较高，能够满足动画中运动物体的实时交互碰撞检测。

关键词: 三角形和三角形相交测试, 碰撞检测, 一类分类, 核函数