pyransac3d.fit_state

FitState Objects

class FitState(TypedDict)

State of a single RANSAC iteration, which every fit(.) gives to its optional callback.

A callback is useful to plot the fitting progress, inspect intermediate results or implement a custom early-stopping criterion. It is called after every iteration and, if it returns a truthy value, the fit stops early and returns the best result found so far. Iterations which sample degenerate points, like 3 collinear points for a plane, are skipped and never reach the callback.

Treat every array as read-only, because they are the same objects used by the fit.

---

iteration: int

Index of the current iteration, starting on 0.

sample_indices: list[int]

Index, in pts, of the points sampled on this iteration.

sample_points: NDArray[np.float64]

The points sampled on this iteration, np.array (S, 3), where S is how many points the shape needs.

model: dict[str, Any]

Candidate shape of this iteration, keyed by the attributes of the shape being fitted, like center and radius for a Sphere.

inliers: NDArray[np.intp]

Index of the points of pts which fit the candidate of this iteration.

best_model: dict[str, Any]

Best candidate found so far, with the same keys of model.

best_inliers: NDArray[np.intp] | list[int]

Index of the points of pts which fit best_model.

is_best: bool

True when the candidate of this iteration became the new best_model.