Detect landmarks in faces

[landmarks, success] = obj.fit(img, faces)
[...] = obj.fit(..., 'OptionName',optionValue, ...)

Input

• img Input image.
• faces Detected faces {[x,y,w,h], ...}.

Output

• landmarks The detected landmark points for each face {{[x,y], ...}, ...}.
• success success flag.

Options for AAM

• Configs Optional runtime parameter for fitting process, only supported for AAM algorithm. A struct-array of the same length as faces with the following fields:
  • R 2x2 rotation matrix. default eye(2,'single')
  • t 2-elements translation vector. default [size(img,2) size(img,1)] / 2
  • scale scaling factor. default 1.0
  • scaleIdx 0-based model scale index (into Scales vector from the constructor options). default 0