nms

plugins.extract.detect.mtcnn.nms(rectangles: ndarray, scores: ndarray, threshold: float, method: str = 'iom') → tuple[ndarray, ndarray]

Apply non-maximum suppression on ROIs in same scale(matrix version)

Parameters:

• rectangles (ndarray) – The [b, l, t, r, b] bounding box detection candidates

• threshold (float) – Threshold for successful match

• method (str) – “iom” method or default. default: “iom”

• scores (ndarray)

Returns:

• rectangles – The [b, l, t, r, b] bounding boxes

• scores – The associated scores for the rectangles

Return type:

tuple[ndarray, ndarray]