You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+3-1Lines changed: 3 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -85,7 +85,7 @@ Watch here the [crop_video help text](Docs/Help/crop_video.txt)
85
85
86
86
_Warning!!!_ The resolution of the output video might differ from the width/height specified in the JSON file. This is due to limitations of some codecs.
87
87
88
-
### Extract Face Mesh
88
+
### Extract Face Data
89
89
90
90
Finds a face in the video and uses MediaPipe to extract landmarks and other head transformation data.
91
91
@@ -113,6 +113,8 @@ The normalization is performed assuming that some of the points at the border of
113
113
Hence, those points are used to compute a "rigid" orthogonal system. The advantage is that we don't need any other MediaPipe module to estimate the rotation of the head.
114
114
The following pic shows the vectors used for the normalization process. It helps understanding the implementation of the `compute_normalization_params()` function.
115
115
116
+
Since v0.4, the tool uses mediapipe 0.10 FaceLandmarker and can also extract AR-Kit complatible blendshapes.
117
+
116
118

0 commit comments