Improved Wood Species Identification Based On Multi-View Imagery of The Three Anatomical Planes [PREPRINT].

The identification of tropical African wood species based on microscopic imagery is a challenging problem due to the heterogeneous nature of the composition of wood combined with the vast number of candidate species. Image classification methods that rely on machine learning can facilitate this identification, provided that sufficient training material is available. Despite the fact that the three main anatomical sections contain information that is relevant for species identification, current methods only rely on the transversal section. Additionally, commonly used procedures for evaluating the performance of these methods neglect the fact that multiple images often originate from the same tree, leading to an overly optimistic estimate of the performance. We introduce a new image dataset containing microscopic images of the three main anatomical sections of 77 Congolese wood species. A dedicated multiview image classification method is developed and obtains an accuracy (computed using the naive but common approach) of 95%, outperforming the singleview methods by a large margin. An in-depth analysis shows that naive accuracy estimates can lead to a dramatic over-prediction, of up to 60%, of the accuracy. Additional images from the non-transversal sections can boost the performance of machine-learning-based wood species identification methods. Additionally, care should be taken when evaluating the performance of machine-learningbased wood species identification methods to avoid an overestimation of the performance.

Creator (organisation)

Silva, Núbia Rosa Da

Deklerck, Victor

Baetens, Jan

Bulcke, Jan Van den

De Ridder, Maaike

Rousseau, Mélissa

Bruno, Odemir Martinez

Beeckman, Hans

Acker, Joris Van

De Baets, Bernard

Verwaeren, Jan

Funder

Funder name	Awards
Fundação de Amparo à Pesquisa do Estado de São Paulo	Grant no. 2011/01523-1 - Grant no. 2011/21467-9 - Grant no. 2014/06208-5
Conselho Nacional de Desenvolvimento Científico e Tecnológico, Brazil	Grant no. 308449/2010-0 - Grant no. 484312/2013-8 - Grant no. 312718/2018-7
Department for Environment, Food and Rural Affairs, UK Government, United Kingdom	World Forest ID project 29084
Belgian Federal Science Policy Office, Belgium	Brain 2.0, grant no. B2/202/P2/SmartwoodID
Centre for International Forestry Research, Indonesia
XI European Development Fund