Institutional Repository of Vilnius University
Semantic segmentation for autonomous vehicles /
Abstract
dc:descriptionSemantic scene segmentation is one of the essential parts of computer vision for autonomous vehicles. It allows each pixel of a camera image to be assigned to a semantic class (e.g., pedestrians, cars, roads, buildings, plants, etc.), making a highly detailed map of the environment. Such accurate scene understanding is essential for obstacle avoidance and decision making in dynamic traffic conditions. In this bachelor's thesis, a literature review was conducted on the usage of novel neural network architectures for the segmentation and classification of street camera data for autonomous vehicles. Later, model parameters' impact experiments were performed to analyze the impact of external and internal variables on architecture performance by training various architectures with different variables. Finally, results were compiled and conclusions were made.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Dobkevičius, Dominykas,
Subjects
dc:subject × 1Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- eng
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD308336589&prefLang=en_US