Inference types and values

Data Structures
struct	peak_inference_result_detection
	The inference detection result. More...
struct	peak_inference_result_classification
	The inference classification result. More...
struct	peak_inference_result_data
	The peak inference result. More...
struct	peak_inference_info
	The peak information struct for CNN details. More...

Typedefs
typedef struct peak_inference *	peak_inference_handle
	The handle to identify an inference instance.
typedef struct peak_inference_result *	peak_inference_result_handle
	The handle to identify an inference result.

Enumerations
enum	peak_inference_type { PEAK_INFERENCE_TYPE_INVALID , PEAK_INFERENCE_TYPE_DETECTION , PEAK_INFERENCE_TYPE_CLASSIFICATION }
	peak inference types More...
enum	peak_inference_preprocessing_mode { PEAK_INFERENCE_PREPROCESSING_MODE_INVALID , PEAK_INFERENCE_PREPROCESSING_MODE_CAFFE , PEAK_INFERENCE_PREPROCESSING_MODE_TENSORFLOW }
	peak inference preprocessing modes. More...

Functions
peak_status	peak_Inference_CNN_Open (const char *path, peak_inference_handle *hInference)
	Loads a cnn file from the file system and returns a handle to the inference instance.
peak_status	peak_Inference_CNN_Close (peak_inference_handle hInference)
	Unloads a cnn instance.
peak_status	peak_Inference_CNN_ProcessFrame (peak_inference_handle hInference, peak_frame_handle hFrame, peak_inference_result_handle *hInferenceHandle)
	Runs an image through a cnn and returns the inference result. This call is blocking and returns once the inference is finished processing.
peak_status	peak_Inference_CNN_Info_Get (peak_inference_handle hInference, peak_inference_info *info)
	Gets detailed info about the loaded CNN.
peak_status	peak_Inference_Result_Get (peak_inference_result_handle hInferenceHandle, peak_inference_result_data *result)
	Get the result from a previous inference call.
peak_status	peak_Inference_Result_Release (peak_inference_result_handle hInferenceHandle)
	Releases a inference result that was previously received from peak_Inference_CNN_ProcessFrame.
peak_status	peak_Inference_Result_Classification_GetList (peak_inference_result_handle hInferenceHandle, peak_inference_result_classification *resultList, size_t *resultCount)
	Get the result entries from a given classification inference result handle.
peak_status	peak_Inference_Result_Detection_GetList (peak_inference_result_handle hInferenceHandle, peak_inference_result_detection *resultList, size_t *resultCount)
	Get the result entries from a given detection inference result handle.
peak_status	peak_Inference_Statistics_Get (peak_inference_handle hInference, peak_inference_statistics *statistics)
	Gets the inference statistics.
peak_status	peak_Inference_Statistics_Reset (peak_inference_handle hInference)
	Resets the inference statistics.
peak_status	peak_Inference_ConfidenceThreshold_Get (peak_inference_handle hInference, uint32_t *threshold)
	Gets the confidence threshold of the inference.
peak_status	peak_Inference_ConfidenceThreshold_GetRange (peak_inference_handle hInference, uint32_t *minThreshold, uint32_t *maxThreshold, uint32_t *incThreshold)
	Gets the allowed range for setting the confidence threshold.
peak_status	peak_Inference_ConfidenceThreshold_Set (peak_inference_handle hInference, uint32_t threshold)
	Sets the confidence threshold for the inference. Lower thresholds include more results.

Detailed Description

Definitions on inference types and values.

Enumeration Type Documentation

peak_inference_preprocessing_mode

enum peak_inference_preprocessing_mode

peak inference preprocessing modes.

Enumerator
PEAK_INFERENCE_PREPROCESSING_MODE_INVALID	Invalid inference preprocessing mode. Use this value for the initialization of variables of type peak_inference_preprocessing_mode.
PEAK_INFERENCE_PREPROCESSING_MODE_CAFFE	Caffe preprocessing mode.
PEAK_INFERENCE_PREPROCESSING_MODE_TENSORFLOW	Tensorflow preprocessing mode.

peak_inference_type

enum peak_inference_type

peak inference types

The inference type specifies the CNN model.

Enumerator
PEAK_INFERENCE_TYPE_INVALID	Invalid inference type. Use this value for the initialization of variables of type peak_inference_type.
PEAK_INFERENCE_TYPE_DETECTION	CNN detection model.
PEAK_INFERENCE_TYPE_CLASSIFICATION	CNN classification model.

Function Documentation

peak_Inference_CNN_Close()

peak_status peak_Inference_CNN_Close

(

peak_inference_handle

hInference

)

Unloads a cnn instance.

Parameters

[in]

hInference

Handle to an inference instance.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_HANDLE hInference is an invalid handle.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_CNN_Info_Get()

peak_status peak_Inference_CNN_Info_Get	(	peak_inference_handle	hInference,
		peak_inference_info *	info
	)

Gets detailed info about the loaded CNN.

Parameters

[in]	hInference	Handle to an inference instance.
[out]	info	Detailed info about the CNN.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_HANDLE hInference is an invalid handle.

PEAK_STATUS_INVALID_PARAMETER info is an invalid pointer.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_CNN_Open()

peak_status peak_Inference_CNN_Open	(	const char *	path,
		peak_inference_handle *	hInference
	)

Loads a cnn file from the file system and returns a handle to the inference instance.

Parameters

[in]	path	Path to the cnn file UTF-8 encoded.
[out]	hInference	Handle to the newly created inference instance.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_PARAMETER path or hInference is an invalid pointer.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_OUT_OF_MEMORY The library is out of memory.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_CNN_ProcessFrame()

peak_status peak_Inference_CNN_ProcessFrame	(	peak_inference_handle	hInference,
		peak_frame_handle	hFrame,
		peak_inference_result_handle *	hInferenceHandle
	)

Runs an image through a cnn and returns the inference result. This call is blocking and returns once the inference is finished processing.

Parameters

[in]	hInference	Handle to an inference instance.
[in]	hFrame	Handle to a frame.
[out]	hInferenceHandle	Handle to an inference result.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_PARAMETER resultHandle is an invalid pointer.

PEAK_STATUS_INVALID_HANDLE hInference or hFrame is an invalid handle.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_OUT_OF_MEMORY The library is out of memory.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_ConfidenceThreshold_Get()

peak_status peak_Inference_ConfidenceThreshold_Get	(	peak_inference_handle	hInference,
		uint32_t *	threshold
	)

Gets the confidence threshold of the inference.

Parameters

[in]	hInference	Handle to an inference instance.
[out]	threshold	Confidence threshold of the inference.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_HANDLE hInference is an invalid handle.

PEAK_STATUS_INVALID_PARAMETER threshold is an invalid pointer.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_ConfidenceThreshold_GetRange()

peak_status peak_Inference_ConfidenceThreshold_GetRange	(	peak_inference_handle	hInference,
		uint32_t *	minThreshold,
		uint32_t *	maxThreshold,
		uint32_t *	incThreshold
	)

Gets the allowed range for setting the confidence threshold.

Parameters

[in]	hInference	Handle to an inference instance.
[out]	minThreshold	Minimum value of allowed confidence threshold.
[out]	maxThreshold	Maximum value of allowed confidence threshold.
[out]	incThreshold	Increment of the confidence threshold.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_HANDLE hInference is an invalid handle.

PEAK_STATUS_INVALID_PARAMETER minThreshold, or maxThreshold, or incThreshold is an invalid pointer.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_ConfidenceThreshold_Set()

peak_status peak_Inference_ConfidenceThreshold_Set	(	peak_inference_handle	hInference,
		uint32_t	threshold
	)

Sets the confidence threshold for the inference. Lower thresholds include more results.

Parameters

[in]	hInference	Handle to an inference instance.
[in]	threshold	Confidence threshold to set.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_HANDLE hInference is an invalid handle.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_Result_Classification_GetList()

peak_status peak_Inference_Result_Classification_GetList	(	peak_inference_result_handle	hInferenceHandle,
		peak_inference_result_classification *	resultList,
		size_t *	resultCount
	)

Get the result entries from a given classification inference result handle.

Provides the current list of result entries for classification

This function implements the Two Staged Data Query principle.

Parameters

[in]	hInferenceHandle	Handle to an inference result.
[out]	resultList	Pointer to a user allocated array buffer to receive the result entry list. If this parameter is NULL, resultCount will contain the current number of result entries. The required size for resultList in bytes is resultCount x sizeof(peak_inference_result_classification)
[in,out]	resultCount	resultList equal NULL: out: minimal number of result entries resultList must be large enough to hold resultList unequal NULL: in: number of result entries resultList can hold out: number of result entries filled by the function

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_PARAMETER resultCount is an invalid pointer.

PEAK_STATUS_INVALID_HANDLE resultHandle is an invalid handle.

PEAK_STATUS_BUFFER_TOO_SMALL resultList is not NULL and the value of *resultCount is too small to receive the expected amount of data.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_Result_Detection_GetList()

peak_status peak_Inference_Result_Detection_GetList	(	peak_inference_result_handle	hInferenceHandle,
		peak_inference_result_detection *	resultList,
		size_t *	resultCount
	)

Get the result entries from a given detection inference result handle.

Provides the current list of result entries for detection

This function implements the Two Staged Data Query principle.

Parameters

[in]	hInferenceHandle	Handle to an inference result.
[out]	resultList	Pointer to a user allocated array buffer to receive the result entry list. If this parameter is NULL, resultCount will contain the current number of result entries. The required size for resultList in bytes is resultCount x sizeof(peak_inference_result_detection)
[in,out]	resultCount	resultList equal NULL: out: minimal number of result entries resultList must be large enough to hold resultList unequal NULL: in: number of result entries resultList can hold out: number of result entries filled by the function

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_PARAMETER resultCount is an invalid pointer.

PEAK_STATUS_INVALID_HANDLE resultHandle is an invalid handle.

PEAK_STATUS_BUFFER_TOO_SMALL resultList is not NULL and the value of *resultCount is too small to receive the expected amount of data.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_Result_Get()

peak_status peak_Inference_Result_Get	(	peak_inference_result_handle	hInferenceHandle,
		peak_inference_result_data *	result
	)

Get the result from a previous inference call.

Parameters

[in]	hInferenceHandle	Handle to an inference result.
[out]	result	Inference result.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_PARAMETER result is an invalid pointer.

PEAK_STATUS_INVALID_HANDLE resultHandle is an invalid handle.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_Result_Release()

peak_status peak_Inference_Result_Release

(

peak_inference_result_handle

hInferenceHandle

)

Releases a inference result that was previously received from peak_Inference_CNN_ProcessFrame.

Parameters

[in]

hInferenceHandle

Handle to an inference result.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_HANDLE resultHandle is an invalid handle.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.

Note

The result handle is no longer valid after the function has returned.

peak_Inference_Statistics_Get()

peak_status peak_Inference_Statistics_Get	(	peak_inference_handle	hInference,
		peak_inference_statistics *	statistics
	)

Gets the inference statistics.

Parameters

[in]	hInference	Handle to an inference instance.
[out]	statistics	Inference statistics for the given hInference.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_HANDLE hInference is an invalid handle.

PEAK_STATUS_INVALID_PARAMETER statistics is an invalid pointer.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.

peak_Inference_Statistics_Reset()

peak_status peak_Inference_Statistics_Reset

(

peak_inference_handle

hInference

)

Resets the inference statistics.

Parameters

[in]

hInference

Handle to an inference instance.

Returns

PEAK_STATUS_SUCCESS Operation was successful; no error occurred.

PEAK_STATUS_INVALID_HANDLE hInference is an invalid handle.

PEAK_STATUS_NOT_INITIALIZED The library is not initialized.

PEAK_STATUS_ERROR An unexpected internal error occurred.