pbio::Capturer Class Reference

Interface object for detecting or tracking of faces in the images or video sequences. More...

#include <Capturer.h>

Public Types
enum	CapturerType { CAPTURER_TYPE_FRONTAL = 0, CAPTURER_TYPE_LEFT_PROFILE = 1, CAPTURER_TYPE_RIGHT_PROFILE = 2 }
	Capturer types (see also RawSample::SampleType). More...
typedef LightSmartPtr < Capturer >::tPtr	Ptr
	Alias for the type of a smart pointer to Capturer.

Public Member Functions
std::vector< RawSample::Ptr >	capture (const RawImage image)
	Capture faces in a given image or video frame. More...
std::vector< RawSample::Ptr >	capture (const unsigned char *data, int data_size)
	Capture faces in a given image or video frame. More...
RawSample::Ptr	manualCapture (const RawImage image, float left_eye_x, float left_eye_y, float right_eye_x, float right_eye_y)
	Create a sample manually with the marked eye positions. More...
RawSample::Ptr	manualCapture (const RawImage image, const std::vector< RawSample::Point > &points)
	Create a sample manually with the marked positions of all points. Used only to adjust the position of points. More...
RawSample::Ptr	manualCapture (const unsigned char *data, int data_size, float left_eye_x, float left_eye_y, float right_eye_x, float right_eye_y)
	Create a sample manually with the marked eye positions. More...
RawSample::Ptr	manualCapture (const unsigned char *data, int data_size, const std::vector< RawSample::Point > &points)
	Create a sample manually with the marked positions of all points. Used only to adjust the position of points. More...
void	resetHistory ()
	Forget all previous frames and faces (only affects trackers).
CapturerType	getType () const
	Get a capturer type. More...
void	setParameter (const std::string parameter, const double value)
	Change the parameter value in runtime. More...

Detailed Description

Interface object for detecting or tracking of faces in the images or video sequences.

Member Enumeration Documentation

enum pbio::Capturer::CapturerType

Capturer types (see also RawSample::SampleType).

Enumerator
CAPTURER_TYPE_FRONTAL	Only frontal-oriented faces will be captured.
CAPTURER_TYPE_LEFT_PROFILE	Only left-profile-oriented faces will be captured.
CAPTURER_TYPE_RIGHT_PROFILE	Only right-profile-oriented faces will be captured.

Member Function Documentation

std::vector< RawSample::Ptr > Capturer::capture ( const RawImage  image )

inline

Capture faces in a given image or video frame.

Parameters

[in]

image

Image or videoframe.

Returns

Vector of captured face samples.

Warning

Gray-scale images are not supported by the objects created with some configuration files.

std::vector< RawSample::Ptr > Capturer::capture ( const unsigned char *  data, int  data_size  )

inline

Capture faces in a given image or video frame.

Parameters

[in]	data	Pointer to the encoded image or videoframe with 8 or 24 bits per pixel. Supported formats are JPEG, PNG, TIF and BMP.
[in]	data_size	Data buffer size in bytes

Returns

Vector of captured face samples.

Warning

Gray-scale images are not supported by the objects created with some configuration files.

Capturer::CapturerType Capturer::getType ( ) const

inline

Get a capturer type.

Returns

Type of the capturer.

RawSample::Ptr Capturer::manualCapture ( const RawImage  image, float  left_eye_x, float  left_eye_y, float  right_eye_x, float  right_eye_y  )

inline

Create a sample manually with the marked eye positions.

Parameters

[in]	image	Image.
[in]	left_eye_x	X coordinate of the left eye.
[in]	left_eye_y	Y coordinate of the left eye.
[in]	right_eye_x	X coordinate of the right eye.
[in]	right_eye_y	Y coordinate of the right eye.

Returns

Captured face sample.

RawSample::Ptr Capturer::manualCapture ( const RawImage  image, const std::vector< RawSample::Point > &  points  )

inline

Create a sample manually with the marked positions of all points. Used only to adjust the position of points.

Warning

For recognition methods 7.7, 6.7, 8.7: samples that are created with this member function will not contain enough information about facial landmarks (even if given points are from the FDA points set), so the FDA algorithm will be rerun during Recognizer::processing. If you use this member function in order to avoid using RawSample::save and store the image on your own, it's recommended to use new serialization functions: RawSample::saveWithoutImage and FacerecService::loadRawSampleWithoutImage.

Parameters

[in]	image	Image.
[in]	points	Vector of points. It should be of the same size and the points should be in the same order as in the vector received from RawSample::getLandmarks from the original sample.

Returns

Captured face sample.

RawSample::Ptr Capturer::manualCapture ( const unsigned char *  data, int  data_size, float  left_eye_x, float  left_eye_y, float  right_eye_x, float  right_eye_y  )

inline

Create a sample manually with the marked eye positions.

Parameters

[in]	data	Pointer to the encoded image or videoframe with 8 or 24 bits per pixel. Supported formats are JPEG, PNG, TIF and BMP.
[in]	data_size	Data buffer size in bytes
[in]	left_eye_x	X coordinate of the left eye.
[in]	left_eye_y	Y coordinate of the left eye.
[in]	right_eye_x	X coordinate of the right eye.
[in]	right_eye_y	Y coordinate of the right eye.

Returns

Captured face sample.

RawSample::Ptr Capturer::manualCapture ( const unsigned char *  data, int  data_size, const std::vector< RawSample::Point > &  points  )

inline

Create a sample manually with the marked positions of all points. Used only to adjust the position of points.

Warning

For recognition methods 7.7, 6.7, 8.7: samples that are created with this member function will not contain enough information about facial landmarks (even if given points are from the FDA points set), so the FDA algorithm will be rerun during the Recognizer::processing. If you use this member function in order to avoid using RawSample::save and store the image on your own, it's recommended to use new serialization functions: RawSample::saveWithoutImage and FacerecService::loadRawSampleWithoutImage.

Parameters

[in]	data	Pointer to the encoded image or videoframe with 8 or 24 bits per pixel. Supported formats are JPEG, PNG, TIF and BMP.
[in]	data_size	Data buffer size in bytes
[in]	points	Vector of points. It should be of the same size, and the points should be in the same order as in the vector received from RawSample::getLandmarks from the original sample.

Returns

Captured face sample.

void Capturer::setParameter ( const std::string  parameter, const double  value  )

inline

Change the parameter value in runtime.

Parameters

[in]	parameter	Parameter name (tag name in the .xml config file).
[in]	value	New parameter value.