Tissue Segmentation in Histopathological Whole-Slide images with Deep Learning

Pryhoda, Oleksandr

Home
→
Students Research & Project Works | Роботи студентів
→
Факультет прикладних наук
→
Бакалаврська програма "Комп'ютерні науки"
→
2019
→
View Item

dc.contributor.author	Pryhoda, Oleksandr
dc.date.accessioned	2024-02-14T15:54:47Z
dc.date.available	2024-02-14T15:54:47Z
dc.date.issued	2019
dc.identifier.citation	Pryhoda, Oleksandr. Tissue Segmentation in Histopathological Whole-Slide images with Deep Learning / Pryhoda, Oleksandr; Supervisor: Dmytro Fishman; Ukrainian Catholic University, Department of Computer Sciences. – Lviv: 2019. – 46 p.	uk
dc.identifier.uri	https://er.ucu.edu.ua/handle/1/4476
dc.language.iso	en	uk
dc.title	Tissue Segmentation in Histopathological Whole-Slide images with Deep Learning	uk
dc.type	Preprint	uk
dc.status	Публікується вперше	uk
dc.description.abstracten	Development of technologies led to the adoption of new digital imaging solutions in pathology field. One such innovation is whole slide imaging, the main purpose of which is digitalizing the whole glass slide with tissue into a high-resolution image. This image is then divided into sections, which are zoomed for further analysis. The main focus of examination is tissue body, but other materials such as debris, dust, and glass are also presented on the slide. In order to focus only on tissue and to make the analysis process more time- and memory-efficient, tissue location on the slide is predefined. Currently, tissue localization procedure is performed by segmentation algorithms based on classical methods of computer vision. These algorithms require manual tuning and might be inaccurate on images with a lot of debris. The issue could be solved with more adaptive methods like deep neural networks. This thesis presents tissue segmentation pipeline based on deep convolutional neural networks. Proposed pipeline showed that deep learning is capable of segmenting tissue as ac- curately as the currently employed approach.	uk