Abstract
This work introduces the Cartwright-Inspired Complex Spectral Pore Descriptor (CICSP), an interpretable hybrid descriptor for quantifying pore-system architecture in segmented two-dimensional rock images. CICSP integrates classical pore morphometry, connected-component topology, skeleton-based network measures, and a complex spectral representation of pore boundaries motivated by coefficient-growth ideas from multivalent analytic function theory. The descriptor is evaluated as a proof of concept using three segmented thin-section images representing fragmented, branched, and coalesced pore architectures. The results show that CICSP distinguishes pore abundance from pore organization. Rock I exhibits low porosity and many dispersed pore components, Rock II displays the strongest skeleton-branching response, and Rock III is dominated by a highly connected pore body containing nearly all pore area. Pore-scale distributions further indicate that these image-level responses emerge from distinct local populations of pore size, compactness, concavity, spectral roughness, and low-order deformation. These findings demonstrate that CICSP converts binary pore masks into fixed-length, geologically interpretable descriptors that capture fragmentation, branching, coalescence, and boundary complexity. The proposed framework provides a compact quantitative representation of two-dimensional pore systems and may support comparative petrographic analysis, facies discrimination, and future machine-learning workflows in digital rock characterization.