This document summarizes and compares several techniques for face recognition using support vector machines (SVMs). It discusses techniques that use 2D and 3D radial geodesics, Gabor features, heuristic and divisional incremental SVM algorithms, feature dimension reduction with DCT/SVM, 2DLDA with SVM, and an incremental SVM learning algorithm combined with clustering. The techniques are evaluated based on their recognition rates on various databases and their pros and cons are outlined. In general, the techniques that combine feature extraction methods like Gabor features or 2DLDA with SVM classification achieve high recognition rates between 86.9-96.18% but have limitations regarding complexity, dependence on image characteristics, or practical implementation.