Journal Article

Simultaneous cell-by-cell recognition and microfibril angle determination in Japanese hardwoods by polarized optical microscopy combined with semantic segmentation

Kita, Yusuke, Yoshinaga, Arata, Setiyobudi, Titis, Sugiyama, Junji, Awano, Tatsuya

2023

Overview

Hardwood species have flourished in the recent evolutionary history of angiosperms and exhibit pronounced diversity in cell anatomy, such as cell arrangement, cell type, ultrastructure, and cellulose microfibril angle. Optimization of the mass transportation and load-bearing functions of wood can be realized by complex combinations of these structural features. To decipher species-by-species structural optimization strategies, multiscale wood cell structures should ideally be measured simultaneously. However, methodological restrictions have hampered progress in this respect. To address this problem, the present study aimed to measure and analyze the cell-by-cell cellulose microfibril angle and anatomy of Japanese hardwood species using polarized optical microscopy-based retardation imaging and deep learning-based semantic segmentation to achieve cell-by-cell microfibril angle imaging in a wide field of view. Microfibril angle imaging revealed that characteristic ultrastructure and microfibril angle distributions were distinguishable depending on the wood species and cell type, such as fibers, axial parenchyma, and vessel elements. The results implied that cell wall thickness and microfibril angle are correlated in fibers or tracheids of certain vesselless and ring-porous species. The combination of microfibril angle imaging and semantic segmentation presents the opportunity to gain insights into optimization strategies in the hierarchical structure of diverse wood species.

• Publisher: Springer Nature B.V
• Subject: Anatomy / Angiosperms / Cellulose / Hardwoods / Image segmentation / Microscopy / Optical microscopy / Optimization / Semantic segmentation / Semantics / Wall thickness