I have dataset of a person , whose images are rotated 90° or 180° from which I have to crop faces and align using mtcnn or normal way ?
In contributed align_dataset_mtcnn.py is not much useful for alignment of faces which are rotated ? How to make images as rotationally invariant to crop and align faces
I have dataset of a person , whose images are rotated 90° or 180° from which I have to crop faces and align using mtcnn or normal way ? In contributed align_dataset_mtcnn.py is not much useful for alignment of faces which are rotated ? How to make images as rotationally invariant to crop and align faces