carefull with photographic reference; camera lenses distort the image based on distance and do not (typically) present anything close to an orthogonal view. If you know enough about the camera ( focal lenght, distance from subject) you can make a blender camera and use the background image for that. look at the front view for instance: the camera is making the ears look smaller and flatter because they’re further away. you’ve increased this more, and you’ve lowered the ears from the reference, probably to match the side view, which also has camera distortion, and may have a slight head tilt. So even though you’ve ‘matched’ the photograph, you are still off. The best way to take a picture is use a huge zoom lense and zoom in from really far away, this’ll flatten out the perspective the most, and still it isn’t perfect. But, using your knowledge of the distortion, correct human proportions, etc, you can allow for the differenses and adjust.
The mesh itself looks like it might give you headaches animating, but hard to tell in edit mode with all the verts visible.