Institutional Repository of Vilnius University
Vaizdų dalinės informacijos panaudojimas jų atpažinime /
Abstract
dc:descriptionThe analysis and classification of image information using neural networks is closely related with image processing techniques. These two techniques are widely used in medicine to study cancer or various eye diseases. The aim of this work is to develop an algorithm which extracts partial information and to investigate its influence on the classification of diabetic retinopathy and pneumonia images datasets. The data used in the study are characterized by imbalances and noise, so for this purpose methods of data balancing and pre-processing have been applied to achieve the main work goal. During the research was noticed, that colored diabetic retinopathy images partial data achieves better results than using black-and-white dataset. The study found that mild and moderate stages of the disease are the most difficult to recognize of all stages of diabetic retinopathy. The classifications of balanced pneumonia images partial information achieved 73.9% accuracy overall, and 98% accuracy in recognizing the disease in a human lung X-ray image. The use of partial image data in the image recognition of medical data has allowed the use of large data sets in neural networks with limited resources and gave significant results.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Raudytė, Rolanda,
- Contributors dc:contributor
-
- Grinis, Irus
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:ELABAETD81707265&prefLang=en_US