Surface digital image correlation techniques are extended to the analysis of volume data obtained from X-ray computed tomography with micrometric resolution or serial reconstruction obtained under plane laser illumination for transparent materials. The local image contrast is provided by markers much smaller than the characteristic dimensions of the gradients to analyze and which do not disturb notably the latter, as checked by numerical simulations. In case of computed tomography, small dense particles are randomly added, while optical scatterers are used in transparent materials. The resolution of the displacement measurement, evaluated on rigid body motions, can be close to 0.01 voxel but can be strongly degraded on noisy images. New matching criteria, less sensitive to noise, are proposed and discussed. Presented applications concern plastic homogeneous flat or notched samples under tension, as well as particulate metal matrix composites.