Work

TTGS: A topology-preserving method for vector reconstruction of plant cell wall networks

Year: 2026

Type: article

Abstract: Plant cell wall architecture influences tissue mechanics and transport, but converting microscopy-derived masks into finite-element geometries requires both accurate centerline localization and preservation of network topology. We developed a topology-preserving thinning-based graph skeletonization method (TTGS) that converts binary cell wall masks into vectorizable node-edge graphs through skeleton extraction, node-cluster identification, path tracing, and connectivity-preserving linearization. Seven reconstruction methods were evaluated using the same binary masks from 12 cross-sectional micrographs of sugarcane ( Saccharum officinarum L. ) parenchyma, vascular bundles, and roots. Across all images, TTGS achieved a structure completeness of 44.04 ± 12.98%, a mean geometric deviation of 2.93 ± 1.04 px, a valid skeleton ratio of 94.49 ± 3.71%, and a breakage rate of 0%. Its unmatched internal endpoint count was 5.00 ± 5.38, while its closed-contour F1 and degree-aware junction F1 reached 93.73 ± 5.56% and 92.75 ± 3.54%, respectively. Among the methods satisfying the predefined short-fragment criterion, TTGS achieved the highest structure completeness and valid skeleton ratio and the lowest geometric deviation. A representative TTGS vector network was imported into finite-element software, reconstructed as a three-dimensional cell-scale model, and successfully solved to obtain a continuous stress distribution. These results show that TTGS balances geometric fidelity with topology preservation and supports the workflow from microscopy-derived masks to vector reconstruction and finite-element analysis.

Source: Plant Phenomics

DOI: 10.1016/j.plaphe.2026.100294