Institutional Repository of Vilnius University
Plaučių echoskopijos naudojimas ligų aptikimui /
Abstract
dc:descriptionLung ultrasound is an innovative method for monitoring lung conditions and diagnosing pathologies. This examination has been of a great help during the Covid19 pandemic, as it demonstrated excellent results in assessing the patient’s condition. Although ultrasound images provide valuable information, their interpretation can be complex. Therefore, deep convolutional networks were applied, that automatically recognize signs of disease and can diagnose pathology. Developing a model with good accuracy that can be applied in the real world is a significant challenge due to the scarcity and imbalance of images. Hence, this work analyzed how different optimizers and loss functions affect the model’s effectiveness when working with lung ultrasound images. The highest accuracy (97%) was achieved by models using the crossentropy loss function with Adam and Nadam optimizers and a learning rate of 0,001. Meanwhile, the RMSProp optimizer with a learning rate of 0,001 and a focal loss function showed the best generalization on new and unseen data.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Simutytė, Rugilė,
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:ELABAETD210647966&prefLang=en_US