Login / Signup

The Outline of Seed Silhouettes: A Morphological Approach to Silene (Caryophyllaceae).

José Javier Martín-GómezJosé Luis Rodríguez-LorenzoÁngel TocinoBohuslav JanousekAna JuanEmilio Cervantes
Published in: Plants (Basel, Switzerland) (2022)
Seed morphology is an important source of information for plant taxonomy. Nevertheless, the characters under study are diverse, and a simple, unified method is lacking in the literature. A new method for the classification of seeds of the genus Silene based on optical images and image analysis has recently been described on the basis of morphological measurements of the lateral seed views. According to the outline of their silhouettes, seeds from 52 species (49 of Silene and three related species) were classified in three groups: smooth, rugose and echinate, revealing remarkable differences between these groups. This methodology has been applied here to 51 new species, making a total of 100 species of Silene analyzed so far. According to our data, a new group was described, termed papillose. The results showed morphological differences between the four mentioned seed groups, with reduced values of circularity for dorsal and lateral seed views in the papillose and echinate groups and reduced values of solidity in the papillose seeds. The method was applied to the analysis of individual as well as to average seed silhouettes and some of the differences between groups were maintained in both cases.
Keyphrases
  • deep learning
  • systematic review
  • machine learning
  • minimally invasive
  • spinal cord
  • healthcare
  • convolutional neural network
  • optical coherence tomography
  • big data
  • genetic diversity
  • health information