The simple reality is that a photo contains no depth information. It is simply light and shadow. We know from our own experience looking at such objects just how it would be constructed in 3D. For example, with a photo of a face, we know that the nose is going to be the closest point to us. The computer lacks this real-life knowledge.
What this means is that there is no substitute for human sculpting. You can apply your image to a grid in ZBrush and even project it onto an object but you will have to sculpt it yourself to interpret the light and shadow into actual depth. Once you have that, ZBrush can then use your sculpt as the source from which to create a depth map – which will be your alpha.