Object Detection Using Deep Neural Models

Ovaj rad opisuje zadatak otkrivanja predmeta kod digitalnih slika primjenom dubokog učenja. Naglasak rada je na opisu popularnog modela YOLOv5. Opisan je postupak učenja takvog modela na vlastito odabranom skupu podataka.This thesis describes the task of object detection in digital images using deep learning. The emphasis of the thesis is on the description of the popular YOLOv5 model. The procedure for learning such a model on a custom data set is described