Open pokoso opened 4 years ago
I don't know the exact details of how dlib is performing landmark detection. OpenFace is using quite a different algorithm, it uses a CE-CLM (Convolutional Expert Constrained Local Model) model, which is a combination of detection of local patches, with optimization over a shape model, such a local and optimization approach typically leads to much less shaking when compared to direct prediction of landmarks.
Hello Tadas,
Compared to the face landmark detection feature of dlib,
The result of landmark points of dlib has some shaking,
but OpenFace has no shaking of landmarks completely.
Can you explain what's different?
Thank you.