Imaging technologies such as CT and MRI generate volumetric (3D) images which reveal the internal structure of opaque objects.

Augmented Reality (AR) devices, such as the Microsoft HoloLens, blend virtual images into the real world directly in a user's field of view.

A volumetric image can be displayed in alignment with an associated physical object to give a user ``x-ray vision''.

The developments required to align volumetric images to objects and visualise them in real-time on an AR device are explored.

A volume renderer which outperforms the leading open-source alternative and can remotely render to AR devices over Wi-Fi has been developed.

Object pose estimation techniques such as GPU-accelerated geometric alignment and marker-based tracking are discussed.

This work could improve outcomes from surgical operations by displaying medical images directly over patients.

Surgeons would not have to look away to an external display which would improve hand-eye coordination, give depth perception and generally improve efficacy over conventional image-guided surgery.