High Accuracy Fundamental Matrix Computation and Its Performance Evaluation

We compare the convergence performance of different numerical schemes for computing the fundamental matrix from point correspondences over two images. First, we state the problem and the associated KCR lower bound. Then, we describe the algorithms of three well-known methods: FNS, HEIV, and renormalization, to which we add Gauss-Newton iterations. For initial values, we test random choice, least squares, and Taubin’s method. Experiments using simulated and real images reveal different characteristics of each method. Overall, FNS exhibits the best convergence performance.

• Type: Journal
• Year: 2007
• Where: IEICET
• Authors: Ken-ichi Kanatani, Yasuyuki Sugaya
• Keywords: Associated Kcr, Convergence Performance, IEICET 2007, Point Correspondences