fsdk::IDetector Struct Reference

face detector interface. More...

#include <IDetector.h>

Inheritance diagram for fsdk::IDetector:

List of all members.

Public Member Functions
virtual ResultValue< FSDKError, Ref< IFaceDetectionBatch > >	detect (Span< const Image > images, Span< const Rect > ROIs, uint32_t perImageNum, DetectionType type=DT_BBOX) noexcept=0
	Detect faces and their parameters on multiple images.
virtual ResultValue< FSDKError, Face >	detectOne (const Image &image, const Rect &rect, DetectionType type=DT_BBOX) noexcept=0
	Light function to get just one best detection from single input image.
virtual ResultValue< FSDKError, Ref< IFaceDetectionBatch > >	redetect (Span< const Image > images, Ref< IFaceDetectionBatch > detectionBatch, DetectionType type=DT_BBOX) noexcept=0
	Batched redetect faces on multiple images. based on the detection results for the previous frames.
virtual ResultValue< FSDKError, Ref< IFaceDetectionBatch > >	redetect (Span< const Image > images, Span< Span< const Detection >> detections, DetectionType type=DT_BBOX) noexcept=0
	Batched redetect faces on multiple images. based on the detection results for the previous frames.
virtual ResultValue< FSDKError, Face >	redetectOne (const Image &image, const Detection &detection, DetectionType type=DT_BBOX) noexcept=0
	Redetect face.
virtual void	setDetectionComparer (DetectionComparerType comparerType) noexcept=0
	Set detection comparer from SDK defined list.
virtual void	setCustomDetectionComparer (const IDetectionComparer *comparer) noexcept=0
	Set custom detection comparer object.
virtual Result< FSDKError >	validate (Span< const Image > images, Span< Span< const Detection >> detections, Span< Span< Result< FSDKError >>> errors) const noexcept=0
	Validate input of multiple frames in a single function call.
virtual Result< FSDKError >	validate (Span< const Image > images, Span< const Rect > rects, uint32_t detectionPerImageNum, Span< Result< FSDKError >> outErrors) const noexcept=0
	Validate input of multiple frames in a single function call.
virtual Result< FSDKError >	validate (Span< const Image > images, Ref< IFaceDetectionBatch > detectionBatch, Span< Result< FSDKError >> outErrors) const noexcept=0
	Validate input of multiple frames in a single function call.
virtual FaceBatchFuture	detectAsync (Span< const Image > images, Span< const Rect > rectangles, uint32_t perImageNum, DetectionType type=DT_BBOX) const =0
	Asynchronously detect faces and their parameters on multiple images.
virtual FaceBatchFuture	redetectAsync (Span< const Image > images, IFaceDetectionBatchPtr detectionBatch, DetectionType type=DT_BBOX) const =0
	Asynchronously redetect faces on multiple images. based on the detection results for the previous frames.

---

Detailed Description

face detector interface.

---

Member Function Documentation

virtual ResultValue<FSDKError, Ref<IFaceDetectionBatch> > fsdk::IDetector::detect	(	Span< const Image >	images,
		Span< const Rect >	ROIs,
		uint32_t	perImageNum,
		DetectionType	type = DT_BBOX
	)		[pure virtual]

Detect faces and their parameters on multiple images.

Parameters:

[in]	images	span of source images.
[in]	rectangles	input rectangles of interest (ROI) span.
[in]	perImageNum	the max number of detections per input image.
[in]	type	type of the detection.

Returns:

ResultValue with error code and IFaceDetectionBatch object.

See also:

Ref, Span, Image, Rect, DetectionType, IFaceDetectionBatch, ResultValue and FSDKError for details.

Note:

images format must be R8G8B8,

See also:

Format.

Note:

all spans should be based on user owned continuous collections.

all spans should be equal size.

virtual FaceBatchFuture fsdk::IDetector::detectAsync	(	Span< const Image >	images,
		Span< const Rect >	rectangles,
		uint32_t	perImageNum,
		DetectionType	type = DT_BBOX
	)		const [pure virtual]

Asynchronously detect faces and their parameters on multiple images.

Parameters:

[in]	images	span of source images.
[in]	rectangles	input rectangles of interest (ROI) span.
[in]	perImageNum	the max number of detections per input image.
[in]	type	type of the detection.

Returns:

Future with IFaceDetectionBatch object.

See also:

Ref, Span, Image, Rect, DetectionType, IFaceDetectionBatch vlc::future for details.

Note:

images format must be R8G8B8,

See also:

Format.

Note:

all spans should be based on user owned continuous collections.

all spans should be equal size.

this method is experimental and interface may be changed in the future.

this method is not marked as noexcept and may throw an exception.

virtual ResultValue<FSDKError, Face> fsdk::IDetector::detectOne	(	const Image &	image,
		const Rect &	rect,
		DetectionType	type = DT_BBOX
	)		[pure virtual]

Light function to get just one best detection from single input image.

Parameters:

[in]	image	source image.
[in]	rect	rectangle of interest in the image.
[in]	type	type of detection: BBox, 5landmarks or 68landmarks.

Returns:

ResultValue with error code and a Face object (detection bbox, landmarks, etc).

See also:

Face, Image, Rect, DetectionType, ResultValue and FSDKError for details.

Note:

image format must be R8G8B8,

See also:

Format.

virtual ResultValue<FSDKError, Ref<IFaceDetectionBatch> > fsdk::IDetector::redetect	(	Span< const Image >	images,
		Ref< IFaceDetectionBatch >	detectionBatch,
		DetectionType	type = DT_BBOX
	)		[pure virtual]

Batched redetect faces on multiple images. based on the detection results for the previous frames.

Parameters:

[in]	images	span of source images.
[in]	detectionBatch	result of detection on the previous frames - Ref with an IFaceDetectionBatch object.
[in]	type	type of redetection.

Returns:

ResultValue with error code and IFaceDetectionBatch object.

See also:

Ref, Span, Image, DetectionType, IFaceDetectionBatch, ResultValue and FSDKError for details.

Note:

images format must be R8G8B8,

See also:

Format.

Note:

all spans should be based on user owned continuous collections.

images span should be the same size with detectionBatch size.

In case if some face from the input detectionBatch was not found the corresponding detection in the output IFaceDetectionBatch object will be invalid.

virtual ResultValue<FSDKError, Ref<IFaceDetectionBatch> > fsdk::IDetector::redetect	(	Span< const Image >	images,
		Span< Span< const Detection >>	detections,
		DetectionType	type = DT_BBOX
	)		[pure virtual]

Batched redetect faces on multiple images. based on the detection results for the previous frames.

Parameters:

[in]	images	span of source images.
[in]	detections	span of detection coordinates in corresponding source images space from the previous frames. It is a two dimensional Span. There is one Span of the rectangles for each image.
[in]	type	type of redetection.

Returns:

ResultValue with error code and IFaceDetectionBatch object.

See also:

Ref, Span, Image, DetectionType, Detection, IFaceDetectionBatch, ResultValue and FSDKError for details.

Note:

images format must be R8G8B8,

See also:

Format.

Note:

all spans should be based on user owned continuous collections.

all spans should be equal size.

If for some of the input detections the redetected face will not be found the appropriate detection in the IFaceDetectionBatch object will be invalid.

virtual FaceBatchFuture fsdk::IDetector::redetectAsync	(	Span< const Image >	images,
		IFaceDetectionBatchPtr	detectionBatch,
		DetectionType	type = DT_BBOX
	)		const [pure virtual]

Asynchronously redetect faces on multiple images. based on the detection results for the previous frames.

Parameters:

[in]	images	span of source images.
[in]	detectionBatch	result of detection on the previous frames - Ref with an IFaceDetectionBatch object.
[in]	type	type of redetection.

Returns:

Future with IFaceDetectionBatch object.

See also:

Ref, Span, Image, DetectionType, IFaceDetectionBatch and vlc::future for details.

Note:

images format must be R8G8B8,

See also:

Format.

Note:

all spans should be based on user owned continuous collections.

images span should be the same size with detectionBatch size.

In case if some face from the input detectionBatch was not found the corresponding detection in the output IFaceDetectionBatch object will be invalid.

this method is experimental and interface may be changed in the future.

this method is not marked as noexcept and may throw an exception.

virtual ResultValue<FSDKError, Face> fsdk::IDetector::redetectOne	(	const Image &	image,
		const Detection &	detection,
		DetectionType	type = DT_BBOX
	)		[pure virtual]

Redetect face.

Parameters:

[in]	image	source image. Format must be R8G8B8.
[in]	detection	detection coordinates in image space from the previous frame to make redetect.
[in]	type	type of detection: BBox, 5landmarks or 68landmarks.

Returns:

ResultValue with error code and Face.

See also:

Face, Image, DetectionType, Detection, ResultValue and FSDKError for details.

Note:

image format must be R8G8B8,

See also:

Format.

virtual void fsdk::IDetector::setCustomDetectionComparer

(

const IDetectionComparer *

comparer

)

[pure virtual]

Set custom detection comparer object.

Parameters:

[in]

comparer

pointer to user defined comparer object.

See also:

IDetectionComparer for details.

Note:

Client code still owns comparer object

virtual void fsdk::IDetector::setDetectionComparer

(

DetectionComparerType

comparerType

)

[pure virtual]

Set detection comparer from SDK defined list.

Parameters:

[in]

comparerType

type of the comparer for detections.

See also:

DetectionComparerType for details.

virtual Result<FSDKError> fsdk::IDetector::validate	(	Span< const Image >	images,
		Span< Span< const Detection >>	detections,
		Span< Span< Result< FSDKError >>>	errors
	)		const [pure virtual]

Validate input of multiple frames in a single function call.

Parameters:

[in]	images	span of source images.
[in]	detections	span of detection coordinates in corresponding source images space from the previous frames. It is a two dimensional Span. There is one Span of Detections for each image.
[out]	errors	output span of errors for each image. It is a two dimensional Span. There is one Span of the errors for each image.

Returns:

Result with error code.

See also:

Span, Image, Detection, Result and FSDKError for details.

Note:

images format must be R8G8B8,

See also:

Format.

Note:

all spans should be based on user owned continuous collections.

all spans should be equal size.

virtual Result<FSDKError> fsdk::IDetector::validate	(	Span< const Image >	images,
		Span< const Rect >	rects,
		uint32_t	detectionPerImageNum,
		Span< Result< FSDKError >>	outErrors
	)		const [pure virtual]

Validate input of multiple frames in a single function call.

Parameters:

[in]	images	span of source images.
[in]	rects	span of rectangle coordinates of corresponding source images.
[in]	detectionPerImageNum	max number of detections per input image.
[out]	errors	output span of errors for each image.

Returns:

Result with error code.

See also:

Span, Image, Rect, Result and FSDKError for details.

Note:

images format must be R8G8B8,

See also:

Format.

Note:

all spans should be based on user owned continuous collections.

all spans should be equal size.

virtual Result<FSDKError> fsdk::IDetector::validate	(	Span< const Image >	images,
		Ref< IFaceDetectionBatch >	detectionBatch,
		Span< Result< FSDKError >>	outErrors
	)		const [pure virtual]

Validate input of multiple frames in a single function call.

Parameters:

[in]	images	span of source images.
[in]	detectionBatch	result of detection on the previous frames - Ref with an IFaceDetectionBatch object.
[out]	errors	output span of errors for each image.

Returns:

Result with error code.

See also:

Ref, Span, Image, IFaceDetectionBatch, Result and FSDKError for details.

Note:

images format must be R8G8B8,

See also:

Format.

Note:

all spans should be based on user owned continuous collections.

all spans should be equal size.

---

The documentation for this struct was generated from the following file:

• include/fsdk/IDetector.h