Login / Signup

Seed Morphology in Silene Based on Geometric Models.

José Javier Martín-GómezAgnieszka RewiczJosé Luis Rodríguez-LorenzoBohuslav JanousekEmilio Cervantes
Published in: Plants (Basel, Switzerland) (2020)
Seed description in morphology is often based on adjectives such as "spherical", "globular", or "reniform", but this does not provide a quantitative method. A new morphological approach based on the comparison of seed images with geometric models provides a seed description in Silene species on a quantitative basis. The novelty of the proposed method is based in the comparison of the seed images with geometric models according to a cardioid shape. The J index is a measurement that indicates the seed percentage of similarity with a cardioid or cardioid-derived figures used as models. The seeds of Silene species have high values of similarity with the cardioid and cardioid-derived models (J index superior to 90). The comparison with different figures allows species description and differentiation. The method is applied here to seeds of 21 species and models are proposed for some of them including S. diclinis, an endangered species. The method is discussed in the context of previous comparison with the measures used in traditional morphometric analysis. The similarity of seed images with geometric figures opens a new perspective for the automatized taxonomical evaluation of samples linking seed morphology to functional traits in endangered Silene species.
Keyphrases
  • deep learning
  • convolutional neural network
  • genetic diversity
  • high resolution
  • dna methylation
  • mass spectrometry
  • clinical evaluation