Digital rock physics (DRP) is a rapidly evolving field of study. One component of digital rock that has not received sufficient attention is how well actual rocks are represented in DRP. Instead, the digital rock community is focused on characterizing the pore space in volumes of rock imaged by microcomputed tomography (micro-CT) and simulating flow through that digitized pore network. This enables computational simulations of routine core analysis measurements, which may be completed in hours instead of days or weeks. Although this alone makes digital rock a worthwhile endeavor, it overlooks much of the detailed textural and compositional information stored within digital rock images below the resolution of micro-CT imaging. This information may be observed in high-resolution 2D transmitted light microscopy images. Textural information impacts not only the tortuosity of the flow path, impacting permeability, but also influences how the rock will respond to stress. Compositional information could also be extracted to not only better characterize the wettability of rocks for relative permeability simulations, but also to supplement petrographic information in diagenetic modeling, among other applications. Ultimately, a full characterization of a digital rock should replicate the acoustic, geomechanical, and petrophysical properties of the imaged sample. The first step toward achieving full digital simulation of rock properties is the fundamental characterization of the sample — extracting the textural and compositional information from digital rock images. Unfortunately, this is a nontrivial undertaking. It involves acquiring sample images, segmenting pores from individual rock minerals, separating these minerals into individual grains and cements, and computing multiple attributes from the segmented grains. To address this issue, we are developing a workflow to compute key textural attributes from images with a long-term vision for the incorporation of geologic characterization into DRP using machine learning.