Показати скорочений опис матеріалу
dc.contributor.author | Petryshak, Bohdan![]() |
|
dc.date.accessioned | 2024-02-14T16:59:26Z | |
dc.date.available | 2024-02-14T16:59:26Z | |
dc.date.issued | 2019 | |
dc.identifier.citation | Petryshak, Bohdan. Medical image segmentation using shape prior information and deep neural networks / Petryshak, Bohdan; Supervisor: Jan Kybic; Ukrainian Catholic University, Department of Computer Sciences. – Lviv: 2019. – 47 p. | uk |
dc.identifier.uri | https://er.ucu.edu.ua/handle/1/4485 | |
dc.language.iso | en | uk |
dc.title | Medical image segmentation using shape prior information and deep neural networks | uk |
dc.type | Preprint | uk |
dc.status | Публікується вперше | uk |
dc.description.abstracten | Semantic image segmentation is the task of classifying each pixel of an image into a corresponding category of what is being represented. It is an essential step to- wards automating image analysis process. However, the low-quality signal, high level of noise, variety of objects appearance, little amount of labeled data are the critical obstacles which stand on the way of achieving the perfect segmentation re- sults. Incorporating the shape prior knowledge has proven significant improvement of the segmentation results. In this work, we extend the existing method of incorpo- rating shape priors within the FCN segmentation framework to a multiclass seman- tic segmentation. We demonstrate the superiority of our extension in five different datasets and show that it capable of making the segmentation results more accurate and plausible in multiclass problems. . . . | uk |