THE POSSIBILITIES OF USING ASK ANALYSIS AND THE INTELLIGENT EIDOS SYSTEM TO IDENTIFY CULTIVARS AND HYBRIDS OF CHRYSANTHEMUM BY THE LEAF CONTOUR
- The authors:
- Yakushina L. G., Kiseleva N. S., Slepchenko N. A.
- Page numbers:
- 29 - 43
- chrysanthemum, cultivar, hybrid, image system identification, leaf, genotype, similarity-difference, generalized images, "Eidos" system, ASK analysis.
An important task in chrysanthemum breeding is to identify various cultivars and hybrids, particularly by the leaf contours, since the contour shape of a particular leaf contains information about the cultivar: both about the genotype and the phenotype. Automated system-cognitive analysis (ASC-analysis) and its software tools, the intelligent system "Eidos" using scanned images of leaves, are proposed as tools for creating a measuring system and its application. The paper shows the results of a quantitative assessment of the external leaf contour in chrysanthemum cultivars and hybrids, on the basis of which generalized images of the studied genotype classes are formed according to the leaf shape; the influence of the values of their genotypic features is studied. Using the most reliable INF4 model, with the integral criterion "Sum of knowledge", the problems of quantitative identification of different cultivars and hybrids by leaf contours were solved. The studied chrysanthemum cultivar samples that belong to the collection of the FRC SSC of RAS are divided into two main clusters, including smaller groups, according to the characteristics of the leaf contour determined by them.