DFX Extraction library v4.14.5
enabling applications to perform DFX blood flow analysis
|
Pose points are used to identify individual MPEG-4 Facial Data Points of a face. More...
#include "dfx/PosePoint.h"
Public Attributes | |
cv::Point3f | point |
the X,Y,Z point location information. More... | |
bool | valid |
if this point is valid, false otherwise. More... | |
bool | estimated |
if this point is "well known", or estimated based on other points. More... | |
float | quality |
the probability quality probability of this point between zero and one. More... | |
Pose points are used to identify individual MPEG-4 Facial Data Points of a face.
DFX utilizes a subset of the 84 feature points described in the MPEG-4 Facial Data Points when constructing regions of interest.
bool dfx::PosePoint::estimated |
if this point is "well known", or estimated based on other points.
When the MPEG-4 feature points are based on machine learning techniques, this value would be false as they are assumed to be more strongly known. On the otherhand, if this point was extrapolated or estimated based on other points that were machine learned, this flag should be set to true.
cv::Point3f dfx::PosePoint::point |
the X,Y,Z point location information.
Presently, the Z is ignored for all study processing as the image is fundamentally 2-D and we are able to construct accurate regions from just the X and Y.
float dfx::PosePoint::quality |
the probability quality probability of this point between zero and one.
Providing this information can help the region definitions chose it's anchor points more accurately.
bool dfx::PosePoint::valid |
if this point is valid, false otherwise.
Because a face maybe partially obscured, not all points may be valid within a specified face and this flag is used to indicate those points which are not valid.