Appendix E. Model Fitting
Several motion-estimation, 3D-structure estimation, and image/motion-segmentation problems require fitting a model to available data samples. This appendix presents model-fitting solutions that are common to these problems. Linear least-squares or total least-squares methods are used when available data samples are free from outliers. Random-sample consensus (RANSAC) is an effective method to deal with data with outliers.
E.1 Least-Squares Fitting
In order to keep the presentation simple, let’s suppose that the available data consists of pairs of points (x1, y1), ..., (xN, yN) and we want to fit a line y = m x + b to these points such that sum of squared vertical distances between the given points and the line δi = ...
Get Digital Video Processing, Second Edition now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
