Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/870CD3C4-73D7-4B52-87DE-B9FFA3BBBBEE | ||
| Year | 2021 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Emmanouil Gialitakis, "Spectral imaging and machine learning technologies for automation of diagnostic microscopy", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2021 https://doi.org/10.26233/heallink.tuc.90000 | ||
| Appears in Collections |
Microscopy for years is an instrumental technology for analyzing tissue samples and locating cancerous cells. The biopsy is a process that can last for several days and is crucial, since the doctor will decide on the most suitable treatment depending on the results. The goal of this thesis is to speed up the process of analyzing a biopsy by using image stitching algorithms. By creating high-resolution mosaics of the samples, it will be easier for different doctors to examine the same sample, while being located in a different area or re-examine the same sample, if needed. The chosen algorithm for stitching is SIFT, which is distinguishable among other algorithms due to its high accuracy. At the same time, the main disadvantage that needs to be mended is the time needed to complete a stitching due to its high complexity. Using various techniques aiming to reduce the elapsed time, like reducing the image size to be analyzed each time and defining regions of interest in the images, can reduce the time needed. By applying those techniques, it is possible to speed up the average time of a single horizontal stitch by approximately four times. These results suggest that it will be viable for the algorithm to be used in a microscope, reducing the time of analysis of the samples.
Copyright © DIAS 2013
