openvino.runtime.opset10.generate_proposals

openvino.runtime.opset10.generate_proposals(im_info: Union[openvino._pyopenvino.Node, int, float, numpy.ndarray], anchors: Union[openvino._pyopenvino.Node, int, float, numpy.ndarray], deltas: Union[openvino._pyopenvino.Node, int, float, numpy.ndarray], scores: Union[openvino._pyopenvino.Node, int, float, numpy.ndarray], min_size: float, nms_threshold: float, pre_nms_count: int, post_nms_count: int, normalized: bool = True, nms_eta: float = 1.0, roi_num_type: str = 'i64', name: Optional[str] = None) openvino._pyopenvino.Node

Return a node which performs GenerateProposals operation.

Parameters
  • im_info – Input with image info.

  • anchors – Input anchors.

  • deltas – Input deltas.

  • scores – Input scores.

  • min_size – Specifies minimum box width and height.

  • nms_threshold – Specifies threshold to be used in the NMS stage.

  • pre_nms_count – Specifies number of top-n proposals before NMS.

  • post_nms_count – Specifies number of top-n proposals after NMS.

  • normalized – Specifies whether proposal bboxes are normalized or not. Optional attribute, default value is True.

  • nms_eta – Specifies eta parameter for adaptive NMS., must be in range [0.0, 1.0]. Optional attribute, default value is 1.0.

  • roi_num_type – Specifies the element type of the third output rpnroisnum. Optional attribute, range of values: i64 (default) or i32.

  • name – The optional name for the output node.

Returns

New node performing GenerateProposals operation.