[專題演講] 【11月13日】陳邦憲 / Geman-McClure Robust Estimation for Spatial Perceptions (Presentations after NCTS Research Abroad Program)

Robust estimation is essential in computer vision, robotics, and navigation, aiming to minimize the impact of outlier measurements for improved accuracy. We present two Geman-McClure robust estimation solvers, FracGM and QGM, leveraging fractional programming and quadratic programming techniques. We demonstrate both solvers on spatial perception benchmark problems, outperforming existing state-of-the-art methods in both accuracy and robustness.