ObjectDetectionAveragePrecision.get_classes_scores_at#

ObjectDetectionAveragePrecision.get_classes_scores_at(res: ndarray, iou: Optional[float] = None, area: Optional[str] = None, max_dets: Optional[int] = None, get_mean_val: bool = True, zeroed_negative: bool = True, class_weights: Optional[ndarray] = None)[source]#

Get the mean value of the classes scores and the result values.

Parameters
res: np.array

either precision or recall when using the ‘2’ return option

ioufloat, default: None

filter by iou threshold

areastr, default: None

filter by are range name [‘small’, ‘medium’, ‘large’, ‘all’]

max_detsint, default: None

filter by max detections

get_mean_valbool, default: True

get mean value if True, if False get per class

zeroed_negativebool, default: True

if getting the class results list set negative (-1) values to 0

class_weightsnp.array, default None

The class weights for weighted macro averaging. If None, gives equal weights to all the classes.

Returns
Union[List[float], float]

The mean value of the classes scores or the scores list.