VAD com.intel.openvx.haar_classifier

[SSchm9]

Hi,

Looking for help  - having an issue with configuring parameters for the com.intel.openvx.haar_classifier. When using VAD, I cannot tell if it is an application issue or configuration issue. I am able to insert the node into, but unsure/unable how to reference my classifier data in parameter (1) via the GUI. I receive the following error when validating the graph:

Status: VX_ERROR_INVALID_TYPE
Operation: vxVerifyGraph
Message: 
Node <node>[1577804979376] (com.intel.openvx.haar_classifier) parameter[2] failed validation
Node <node>[1577804979376] (com.intel.openvx.haar_classifier) failed validation

Attached is a screen shot.

Questions:

1. What is the best way to debug and obtain more descriptive logs when diagnosing an issue?
2. What is the method to add haar classifiers via VAD?

Thanks

[Anna_B_Intel]

Hi Stephen, 

You need to correct the typo at the in KernelsParameters.xml, Line number 392, column 33: please type VX_TYPE_FLOAT instead of VX_TYPE_FLAOT.

Best wishes, 

Anna

[SSchm9]

Hi Anna,
And thanks for the fix. It appears the updates to the KernelsParameters.xml has resolve the initial issue, but now I am receiving an error on parameter 3 (Scalar) which is looking for - num "The number of possible outcomes"... It is not clear in documentation what SCALARTYPE is needed or if the actual "value" is a dependency to validate the node... 

In testing it appears all types listed return the error quoted below, but when choosing VX_TYPE_UINT32, the VAD tool crashes... Is there any significance to that type?

Current Error:

Status: VX_ERROR_INVALID_TYPE
Operation: vxVerifyGraph
Message: 
Node <node>[2237097780048] (com.intel.openvx.haar_classifier) parameter[3] has invalid type
Node <node>[2237097780048] (com.intel.openvx.haar_classifier) parameter[3] failed validation

Please advice..
Stephen