Abstract
dc:descriptionThe work examines the plant characteristics outlined in Povilas Snarskis's book "Guide to Knowing the Plants of the LSSR" that allow for distinguishing plant species, how they can be singled out from digitized flower data, and what digital recognition methods would allow them to be interpreted in solving a classification problem. This work analyzes flower characteristics: physical dimensions of flowers, symmetry, shape, color, and texture. Recognition methods that mimic current human cognition and learning comprehension were selected to analyze features: neural networks, convolutional neural networks, and vision transformers. Based on the acquired theoretical basis, it is demonstrated that neural networks, convolutional neural networks, and vision transformers are potential solutions to the flower species recognition problem. Neural networks of various architectures are trained to solve the classification task, and their classification accuracies are analyzed. The classification accuracy of individual models is improved by combining them into neural network ensembles.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Stoškus, Tomas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD192846000&prefLang=en_US