face_video_segment
0.8
Adaptation of The Video Segmentation Project for face segmentation.
|
Divides a sequence of frames containing faces into keyframes. More...
#include <keyframer.h>
Classes | |
struct | FaceData |
struct | Keyframe |
Public Member Functions | |
Keyframer (int start_frame=10, int stability_range=5) | |
Constructor. More... | |
void | addFrame (const sfl::Frame &sfl_frame, Frame &fvs_frame) |
Add a new frame. More... | |
Divides a sequence of frames containing faces into keyframes.
Each keyframe is associated with it's corresponding face orientation. A new frame will only become a keyframe if the euclidean distance between it's face orientation (in euler angles) to the orientation of any other keyframe is large enough.
fvs::Keyframer::Keyframer | ( | int | start_frame = 10 , |
int | stability_range = 5 |
||
) |
Constructor.
[in] | start_frame | The frame index to start adding keyframes from. |
[in] | stability_range | Specify how many frames in a row must have landmarks before a new keyframe can be added. |
void fvs::Keyframer::addFrame | ( | const sfl::Frame & | sfl_frame, |
Frame & | fvs_frame | ||
) |
Add a new frame.
[in] | sfl_frame | Landmarks frame. |
[in,out] | fvs_frame | Regions frame. |