In summary, the conversation discussed the possibility of reconstructing a whole head using a 3D model created through photogrammetry. However, it was noted that medical technology may not be advanced enough to achieve this. The concept of creating a 3D object from a finite set of plane projections was also mentioned, but it was clarified that this would not result in a unique reconstruction.
Medical advances might not be up to the challenge of reconstructing a whole head.
But if you mean creating a 3D model of the surface of the head (the surface of the head, but nothing below that), you should research the topic of photogrammetry. You'll need more than just a couple of photographs though, and the photographs should ideally be taken in a controlled setting, with known angles and distances to the camera as well as known angles (and perhaps distances) to light sources. This area of technology has advanced quite a bit in recent years.
I recall reading in some recreational mathematics book (it may have been Godel, Escher, Bach) that one can prove the following: given any finite set of plane figures on planes at various angles, there exists a 3-dimensional object whose projections onto those planes are exactly those figures. As a corollary, this 3d object cannot be unique; to any finite set of plane projections, we can always add more plane projections, and construct an object to satisfy the original projections and the new ones. By choosing different new projections, one must have different objects satisfying the original set.
So strictly speaking, no, you cannot reconstruct a 3d figure from a finite collection of projections onto planes. However, with additional assumptions (such as continuity), you might be able to.