create_dl_layer_zoom_factor🔗
Short description🔗
create_dl_layer_zoom_factor — Create a zoom layer using size factors.
Signature🔗
create_dl_layer_zoom_factor( dl_layer DLLayerInput, string LayerName, number ScaleWidth, number ScaleHeight, string Interpolation, string AlignCorners, attribute.name GenParamName, attribute.value GenParamValue, out dl_layer DLLayerZoom )
Description🔗
The operator create_dl_layer_zoom_factor creates a zoom layer using
size factors and returns the layer handle in DLLayerZoom.
The parameter DLLayerInput determines the feeding input layer and
expects the layer handle as value.
The parameter LayerName sets an individual layer name.
Note that if creating a model using create_dl_model each layer of
the created network must have a unique name.
The parameters ScaleWidth and ScaleHeight specify the ratio
between the output and the corresponding input dimension. Together they
define the output size of the zoom layer DLLayerZoom.
The parameter Interpolation defines the interpolation mode.
Currently the modes 'bilinear' and 'nearest_neighbor' are
supported.
The parameter AlignCorners defines how coordinates are transformed
from the output to the input image. Alternatively, the generic parameter
'coordinate_transformation_mode' can be specified. Note, that the
specification of both parameters may not contradict each other:
-
AlignCorners= 'true' or'coordinate_transformation_mode' = 'align_corners': The transformation is applied in the HALCON Non-Standard Cartesian coordinate system (edge-centered, with the origin in the upper left corner, see chapter Transformations / 2D Transformations). Using the \(x\) axis as an example, this leads to:\[\begin{eqnarray*} x_{input} = x_{output} * (length_{input} - 1) / (length_{output} - 1) \end{eqnarray*}\] -
AlignCorners= 'false':-
'coordinate_transformation_mode' ='half_pixel' (default): The transformation is applied in the HALCON standard coordinate system (pixel centered, with the origin in the center of the upper left pixel, see chapter Transformations / 2D Transformations). Using the \(x\) axis as an example, this leads to:
\[\begin{eqnarray*} x_{input} = (x_{output} + 0.5) * length_{input} / length_{output} - 0.5 \end{eqnarray*}\] -
'coordinate_transformation_mode' = 'pytorch_half_pixel':
\[\begin{eqnarray*} x_{input} = length_{output} > 1 ? (x_{output} + 0.5) * length_{input} / length_{output} - 0.5 : 0.0 \end{eqnarray*}\] -
'coordinate_transformation_mode' = 'asymmetric': The transformation is applied asymmetrically. Using the \(x\) axis as an example, this leads to:
\[\begin{eqnarray*} x_{input} = x_{output} * length_{input} / length_{output} \end{eqnarray*}\]
-
The following generic parameters GenParamName and the corresponding
values GenParamValue are supported:
-
'coordinate_transformation_mode': Defines how coordinates are transformed from the output to the input image. If not specified, the parameter
AlignCornersdetermines the transformation.Default: 'half_pixel'
-
'nearest_mode': Defines how to get the “nearest” pixel in the input image. This parameter only applies if
Interpolationis set to 'nearest_neighbor'.Suggested values: 'round_prefer_floor', 'round_prefer_ceil', 'floor', 'ceil'
Default: 'round_prefer_floor'
-
'is_inference_output': Determines whether
apply_dl_modelwill include the output of this layer in the dictionaryDLResultBatcheven without specifying this layer inOutputs('true') or not ('false').Default: 'false'
Certain parameters of layers created using this operator
create_dl_layer_zoom_factor can be set and retrieved using
further operators.
The following tables give an overview, which parameters can be set
using set_dl_model_layer_param and which ones can be retrieved
using get_dl_model_layer_param or get_dl_layer_param.
Note, the operators set_dl_model_layer_param and
get_dl_model_layer_param require a model created by
create_dl_model.
| Layer Parameters | set |
get |
|---|---|---|
'align_corners' (AlignCorners) |
x |
x |
| 'coordinate_transformation_mode' | x |
x |
'input_layer' (DLLayerInput) |
x |
|
'interpolation_mode' (Interpolation) |
x |
|
'name' (LayerName) |
x |
x |
| 'nearest_mode' | x |
x |
'output_layer' (DLLayerZoom) |
x |
|
'scale_params' (ScaleWidth and ScaleHeight) |
x |
|
| 'shape' | x |
|
| 'type' | x |
| Generic Layer Parameters | set |
get |
|---|---|---|
| 'is_inference_output' | x |
x |
| 'num_trainable_params' | x |
Execution information🔗
Execution information
-
Multithreading type: reentrant (runs in parallel with non-exclusive operators).
-
Multithreading scope: global (may be called from any thread).
-
Processed without parallelization.
Parameters🔗
DLLayerInput (input_control) dl_layer → (handle)
Feeding layer.
LayerName (input_control) string → (string)
Name of the output layer.
ScaleWidth (input_control) number → (real / integer)
Ratio output/input width of the layer.
Default: 2.0
ScaleHeight (input_control) number → (real / integer)
Ratio output/input height of the layer.
Default: 2.0
Interpolation (input_control) string → (string)
Mode of interpolation.
Default: 'bilinear'
List of values: 'bilinear', 'nearest_neighbor'
AlignCorners (input_control) string → (string)
Type of coordinate transformation between output/input images.
Default: 'false'
List of values: 'false', 'true'
GenParamName (input_control) attribute.name(-array) → (string)
Generic input parameter names.
Default: []
List of values: 'coordinate_transformation_mode', 'is_inference_output', 'nearest_mode'
GenParamValue (input_control) attribute.value(-array) → (string / integer / real)
Generic input parameter values.
Default: []
Suggested values: 'align_corners', 'asymmetric', 'pytorch_half_pixel', 'half_pixel', 'true', 'false', 'ceil', 'floor', 'round_prefer_ceil', 'round_prefer_floor'
DLLayerZoom (output_control) dl_layer → (handle)
Zoom layer.
Combinations with other operators🔗
Module🔗
Deep Learning Professional