your pictures are not good for 3d reconstruction:
-high compression artifacts, try to use raw images exported with no destructive compression or lowest compression level
-bad lighting, try to have homogeneous lighting all over you face without shadows
-very short depth of field hence very narrow detailed band and a lot of smooth areas, try to use higher camera aperture (and more light) and get the camera further away from your face
in addition, you need more than two images to get a complete facial mask as each point needs at least two cameras to be reconstructed
good luck