mlxtend version: 0.14.0dev
extract_face_landmarks
extract_face_landmarks(img, return_dtype=
Function to extract face landmarks.
Note that this function requires an installation of the Python version of the library "dlib": http://dlib.net
Parameters
-
img
: array, shape = [h, w, ?]numpy array of a face image. Supported shapes are - 3D tensors with 1 or more color channels, for example, RGB: [h, w, 3] - 2D tensors without color channel, for example, Grayscale: [h, w] return_dtype: the return data-type of the array, default: np.int32.
Returns
-
landmarks
: numpy.ndarray, shape = [68, 2]A numpy array, where each row contains a landmark/point x-y coordinates.
Examples
For usage examples, please see http://rasbt.github.io/mlxtend/user_guide/sources/image/extract_face_landmarks.ipynb