RANSAC

RANSAC algorithm used when there are more points than algorithm requires.

Arguments:

• sample_selector: Function that given set of ids of the size n_samples, returns input for the kernel method.
• kernel: Function that calculates candidates from the sampled data.
• rank: Function that evaluates candidates, computed by kernel. It should return n_inliers, model, where model can be of any type. And is not used by the ransac itself. It only uses n_inliers to select which model is better.
• n_points: Total number of points in the data.
• n_samples: Number of points, that kernel function accepts. E.g. five_point_ransac sets this value to 5.
• iterations: Maximum number of iterations to attempt. Default is 100.
• confidence: Confidence in [0, 1] range. The higher the value, the more certain algorithm is that the picked model does not contain outliers. Number of iterations performed by RANSAC depends on this value as well.

Returns:

Number of inliers and the selected model in the format returned by kernel.