Back to search

Institutional Repository of Vilnius University

Mikroplastikų segmentavimas mikroskopiniuose vaizduose /

Abstract

dc:description

This work examined the problem of segmenting microplastics in microscopic images using deep learning-based image segmentation models. U-Net, LinkNet and YOLOv11m-seg models, their operating principles and application possibilities for microplastic analysis were analyzed during the work. In the experimental part, the models were trained using a dataset of microscopic microplastic images and evaluated according to the IoU and Dice coefficient metrics. The best results were achieved by the U-Net model, which segmented microplastic areas most accurately. The LinkNet model demonstrated slightly worse results, and the YOLOv11m-seg model performed the worst in this task. The study also evaluated the impact of data augmentation on the performance of the models, but no significant improvement in segmentation quality was observed. The results showed that semantic segmentation architectures are more suitable for segmenting microplastics in microscopic images than instance segmentation models. The study revealed that the performance of the models strongly depends on the specifics of the dataset and the properties of the segmented objects, and proved that deep learning methods can be effectively applied to automated microplastic analysis.

Degree

thesis:*
Grantor dc:publisher
Institutional Repository of Vilnius University
Year dc:date
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Židonis, Justinas,

Rights

dc:rights
Statement dc:rights
  • info:eu-repo/semantics/openAccess
Language dc:language
lit

Identifiers

dc:identifier.*

Chain of custody

source
Harvested from
Vilnius University
Base URL
epublications.vu.lt/oai
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
related terms
citation

Židonis, Justinas,. Mikroplastikų segmentavimas mikroskopiniuose vaizduose /. Institutional Repository of Vilnius University, 2026. https://repository.vu.lt/VU:ELABAETD308339355&prefLang=en_US