public CogTransform2DLinear StartPose { get; set; }

public CogTransform2DLinear StartPose { get; set; }

Public Property StartPose As CogTransform2DLinear 
	Get 
	Set

Public Property StartPose As CogTransform2DLinear 
	Get 
	Set

public:
property CogTransform2DLinear^ StartPose {
	CogTransform2DLinear^ get ();
	void set (CogTransform2DLinear^ value);
}

public:
property CogTransform2DLinear^ StartPose {
	CogTransform2DLinear^ get ();
	void set (CogTransform2DLinear^ value);
}

Property Value

This property is for advanced users only.

The StartPose is used to specify the expected location of the pattern origin within the input image's selected coordinate space. This is where SearchMax will begin searching for the pattern.

This means that the run parameter DOFs are applied relative to this StartPose property. This makes the start pose a convenient way of varying the runtime DOFs. To illustrate, if the run parameters specify an angle range of (-pi/2, +pi/2) degrees, then with an identity start pose the actual search range will still be (-pi/2, +pi/2) degrees. But if the start pose has a pi/2 degree angle, then the actual search range is effectively (0, pi).

In addition, StartPoseTranslationUncertaintyX and StartPoseTranslationUncertaintyY can be used to limit or expand the range of translation in which the origin can be found. This can be useful in situations where the feature of interest's origin is limited to a small range, but the scale may vary.

Note, however, that the train time DOFs are applied relative to the train client coordinates (selected space of the train image) and pattern Origin, not the StartPose.

Also note that the result pose is reported relative to the run client coordinates (selected space of the run image), not the StartPose.

Rephrased a slightly different way, the "effective" run-time DOFs are made up by combining the DOFs in run parameters with the StartPose. Similarly, the effective train-time DOFs are the made up by combining the DOFs in the train parameters with the pattern’s Origin. This means that the start pose and the pattern origin must be considered when judging whether the train-time DOF ranges include that of the run-time DOF ranges.

Note If the TrainMode is set to CreateDOFTemplates the tool also requires that the run-time DOF ranges based on the pixel space of the run-time image should be included in the train-time DOF range based on the pixel space of the train-time image. This means that if the selected space of the train image or run image is not the pixel space and are not the same, it might happen that the DOF ranges defined in the selected space of the run-time image are included inside the DOF ranges defined in the selected space of the train-time image, but run-time DOFs are outside the train-time DOF ranges in pixel space. SearchMax cannot run under these circumstances because it has not trained the appropriate template pixels to perform the search. This happens mainly when a fixture tool is used before SearchMax tool (selected space of the run image is from the fixture tool). When this happen SearchMax will throw an exception with error message to indicate that it was not trained for the DOF range in pixel space. When this happens users must either train a wider range of DOFs than they intend to use at runtime or set CogSearchMaxPattern to EvaluateDOFsAtRuntime if the runtime client coordinates will vary from the training client coordinates.