University of Twente Student Theses

Login

Transfer learning for cell image reconstruction

Raureanu, A.S. (2021) Transfer learning for cell image reconstruction.

[img] PDF
1MB
Abstract:Whole slide scanners are a useful tool for doing cell analysis in an efficient manner. A big problem with this is that scanning is not always perfect, which results in artifacts such as blur in some parts of the scan. This causes further problems for the specialists using the scanner, as they have to manually inspect the blurry areas in question and give an objective conclusion. To solve this issue, two Convolutional Autoencoders models are designed and implemented to reconstruct cell slide images. The performance of the models to remove blur from sections of cell slides will then be investigated. The robustness of the Autoencoders is also tested on cell images generated artificially that have had Gaussian blur applied to them. Both trained models successfully deblur cell images with minor performance decreases when the blur is caused by the camera lens focusing below the focal plane. The reconstructions of synthetic cells is also achievable with only a 15% performance decrease when deblurring cell images with high amounts of Gaussian blur applied to them.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science BSc (56964)
Link to this item:https://purl.utwente.nl/essays/87001
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page