openvino.runtime.opset10.interpolate#

openvino.runtime.opset10.interpolate(image: Node | int | float | ndarray, output_shape: Node | int | float | ndarray, scales: Node | int | float | ndarray, mode: str, shape_calculation_mode: str, pads_begin: List[int] | None = None, pads_end: List[int] | None = None, coordinate_transformation_mode: str = 'half_pixel', nearest_mode: str = 'round_prefer_floor', antialias: bool = False, cube_coeff: float = -0.75, axes: Node | int | float | ndarray | None = None, name: str | None = None) → Node#

Perform interpolation of independent slices in input tensor.

Parameters:

  • image – The node providing input tensor with data for interpolation.

  • output_shape – 1D tensor describing output shape for spatial axes.

  • scales – 1D tensor describing scales for spatial axes.

  • mode – Specifies type of interpolation. Possible values are: nearest, linear, linear_onnx, cubic.

  • shape_calculation_mode – Specifies which input, sizes or scales, is used to calculate an output shape.

  • pads_begin – Specifies the number of pixels to add to the beginning of the image being interpolated. Default is None.

  • pads_end – Specifies the number of pixels to add to the end of the image being interpolated. Default is None.

  • coordinate_transformation_mode – Specifies how to transform the coordinate in the resized tensor to the coordinate in the original tensor. Default is “half_pixel”.

  • nearest_mode – Specifies round mode when mode == nearest and is used only when mode == nearest. Default is “round_prefer_floor”.

  • antialias – Specifies whether to perform anti-aliasing. Default is False.

  • cube_coeff – Specifies the parameter a for cubic interpolation. Default is -0.75.

  • axes – 1D tensor specifying dimension indices where interpolation is applied. Default is None.

  • name – Optional name for the output node. Default is None.

Returns:

Node representing interpolation operation.