Segmentace volumetrických dat pomocí neuronových sítí

Abstract

This work deals with the reconstruction of the orbital bone from volumetric data to facilitate fracture detection. Orbital bone fractures are a crucial step in diagnosing head and facial injuries. Unfortunately, detecting these fractures can be challenging as the orbital bones are thin and prone to various types of injuries, leading to poor visualization on imaging technologies such as CT and MRI. In addition, insufficient experience of radiologists or physicians analyzing imaging data can lead to missed or incorrectly interpreted fractures. Advanced machine learning and artificial intelligence algorithms have recently been used to address this problem, allowing for precise detection and localization of orbital fractures. C++ and Python with additional libraries will be used to solve this problem.