Završni rad pod mentorstvom prof. dr. sc. Siniše Šegvića na Fakultetu elektrotehnike i računarstva u Zagrebu u akademskoj godini 2019./2020.

Semantička segmentacija kolničkih trakova provedena je nad skupom podataka LLAMAS. Dostupne su jupyter bilježnice koje su korištene, ali i python datoteke istog sadržaja. Sam skup podataka je javno dostupan za preuzimanje na https://unsupervised-llamas.com/llamas/index. Korištena je biblioteka PyTorch, a sama evaluacija i treniranje je provedeno na platformi Google Colaboratoy uz korištenje GPU NVIDIA TESLA K80.

Bachelor thesis under the mentorship of prof. Siniša Šegvić, PhD at the Faculty of Electrical Enigneering and Computing in the academic year 2019/2020

Semantic segmentation of road lanes was conducted on the LLAMAS dataset (Summary and files available on https://unsupervised-llamas.com/llamas/index). Jupyter notebooks used to train the model are provided, as well as the same code in .py format. Model was implemented using PyTorch library and Python 3.7. Evaluation and training were conducted on the Google Colaboratoy platform (GPU NVIDIA TESLA K80).