|
- 2018
Classification of CITES-listed and other neotropical Meliaceae wood images using convolutional neural networksDOI: 10.1186/s13007-018-0292-9 Keywords: Wood identification, Illegal logging, CITES, Forensic wood anatomy, Deep learning, Transfer learning, Convolutional neural networks Abstract: Expected identification relationships based on the generalized wood anatomical distinctness of each group of species (increasing distinctness along the vertical axis) and relative variability within each group of species (variability increasing with increasing bar length along the horizontal axis). The blue tree (confusion cladogram) to the right of the images indicates the expected nested sets of woods likely to be confused with each other based on their anatomical distinctness and variability. Conventional wisdom in wood anatomical identification does not predict species-level resolutio
|