This paper proposes a multispectral feature fusion framework for multispectral wood identification using spectral data acquired by near-infrared (NIR) spectroscopy, hyperspectral imaging (HSI), and terahertz (THz) spectroscopy. For five wood species, the framework combines spectral preprocessing with a three-stage filter–wrapper–embedded feature selection strategy, retaining 13 HSI, 9 NIR, and 22 THz features. To exploit complementary spectral information, this paper develops weighted feature fusion and a modality-attention Transformer fusion (MATF) model incorporating modality-level attention for adaptive cross-modal interaction. Weighted fusion achieved a mean classification accuracy (ACC) of 92.6% ± 4.6% and a mean macro-averaged F1 (Macro-F1) of 92.5% ± 4.7%. Under stratified five-fold cross-validation repeated three times, MATF achieved a mean ACC of 94.46% ± 4.98% and a mean Macro-F1 of 94.43% ± 5.19%. An ablation study showed that removing the Transformer reduced mean ACC and Macro-F1 to 92.13% and 91.59%, respectively, supporting its contribution to classification performance. The framework further incorporates modality-level Shapley additive explanations (SHAP) to interpret the attribution of each spectral source to model predictions. These results support the effectiveness of combining complementary spectral information with adaptive cross-modal interaction for wood species identification.
Wang, Jinsong, Wang, Yuan, Avramidis, Stavros. An interpretable modality-attention transformer architecture for multispectral feature fusion in automated wood species identification. // Wood Science and Technology. – 2026. – DOI: 10.1007/s00226-026-01826-y. – URL: http://link.springer.com/openurl/fulltext?id=doi:10.1007/s00226-026-01826-y
Wang, Jinsong, Wang, Yuan, Avramidis, & Stavros (2026). An interpretable modality-attention transformer architecture for multispectral feature fusion in automated wood species identification. Wood Science and Technology. https://doi.org/10.1007/s00226-026-01826-y
Wang, et al. “An interpretable modality-attention transformer architecture for multispectral feature fusion in automated wood species identification.” Wood Science and Technology, 2026. doi:10.1007/s00226-026-01826-y.
@article{Wang2026,
title = {An interpretable modality-attention transformer architecture for multispectral feature fusion in automated wood species identification},
author = {Wang and Jinsong and Wang and Yuan and Avramidis and Stavros},
journal = {Wood Science and Technology},
year = {2026},
doi = {10.1007/s00226-026-01826-y},
url = {http://link.springer.com/openurl/fulltext?id=doi:10.1007/s00226-026-01826-y}
}
The misrepresentation of timber provenance undermines global supply chain transparency and sustainable forest management practices. Although reference forensic approaches, such as stable isotope analysis and chemical and genetic...
Plant polyphenols from natural products exhibit potent inhibitory effects on two metabolic enzymes, α-amylase and α-glycosidase. Lignin, a natural polymer second only to cellulose in natural abundance, is rich in polyphenolic...
X-ray densitometry was evaluated for its suitability to characterize density profiles in chemically modified wood, using sorbitol–citric acid (SorCA)-modified Scots pine sapwood as a model material. Gravimetric layer-by-layer profiling...
Thermal modification improves wood dimensional stability and can enhance biological durability, but these benefits are often accompanied by losses in strength and fracture resistance. This review examines how treatment conditions govern...
The wetting behavior of bark surfaces is governed by both surface chemistry and three-dimensional geometry, yet how these properties evolve concurrently during weathering remains poorly understood. This study investigated the dynamic...