src.runtime.utils.evaluation namespace

Submodules

src.runtime.utils.evaluation.lane module

class src.runtime.utils.evaluation.lane.LaneEval

Bases: object

static bench(pred, gt, y_samples)

bench one sample

Parameters

• pred – predicted lanes

• gt – validation lanes

• y_samples – y coordinates

Returns: accuracy, false positives, false negatives

static bench_one_submit(json_pred, json_gt)

Do bench for a list of json-results by calling bench() for every sample

Parameters

• json_pred – predicted list

• json_gt – compare list

Returns: bench result

static get_angle(xs, y_samples)

static line_accuracy(pred, gt, thresh)

lr = LinearRegression()

pixel_thresh = 20

pt_thresh = 0.85