Institutional Repository of Vilnius University
Neuroninių tinklų taikymas mikroplastikų segmentavimui ir klasifikavimui /
Abstract
dc:descriptionMicroplastics pose a threat not only to the environment, animals, but also to humanity. Microplastics poison the human body and cause very serious health issues. By recognizing the types of microplastics and classifying them, it is then possible to conduct various microplastic studies and help assess the impact of microplastics not only on ecosystems, but also on humanity. In this bachelor’s thesis, a literature analysis was conducted on the application of neural networks for the segmentation and classification of microplastics. And using the unlabeled dataset provided by the Physical and Technological Science Center (FTMC), a UNet neural network was created, trained and optimized to segment microplastic particles and a pretrained VGG16 neural network was used to perform feature extraction to prepare the logistic regression classifier, together with pseudolabels and clustering, which managed to classify the microplastics. The UNet model accurately segments microplastics in an inhomogeneous background, solving the problem of unlabeled datasets with pseudolabels and clustering, classifier classified microplastics.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Jonaitis, Kristupas,
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:ELABAETD239030812&prefLang=en_US