Institutional Repository of Vilnius University
Automatinės stuburo balansavimo sistemos kūrimas /
Abstract
dc:descriptionCreation of a Backbone Balancing System Spinal disease detection diagnostics, especially AIS (Adolescent idiopathic Scoliosis), is based on international and widely recognized methodology (Lenke classification), as set out in the book "Radiographic Measurement Manual, editors Michael F. O'Brien, Timothy R. Kukla, Kathy M. Blonk, Lawrence G. Lenke” and applied in hospitals. One of the possible ways of diagnosis - spinal identification and its curvature calculations. Proper algorithm to solve this task is presented in [MJS + 14]. The goal of this “Creation of a Backbone Balancing System” bachelor theses is to improve [MJS +14] algorithm by increasing its credibility and to adapt it to different types of x-ray images. During this research the automatic model was designed, created and tested. This model can recognize the type of X-ray image (pelvic or lung area) and to find specific points. For image type identification AlexNet neural network [KSH12] was used (available on MatConvNet platform). It was already trained using ImageNet database. During research this neural network was adjusted using fine-tuning method to recognize only two types of images. Lungs and pelvic bone x-ray images were taken from open source OPENI and MEDPIX databases. Level Set method [LXG +10] was used for points identification. Research input data - X-ray images which are daily used in medical practice.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2017
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Danytė, Ingrida,
- Contributors dc:contributor
-
- Juozapavičius, Algimantas
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:ELABAETD23218292&prefLang=en_US