Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/A33F79FD-4001-4821-9718-D61E8509336A | ||
| Year | 2019 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Sevastianos Papadogiannis, "Bio-optical modeling of epithelial neoplasia", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2019 https://doi.org/10.26233/heallink.tuc.83622 | ||
| Appears in Collections |
Over the last years, a radical change regarding diagnosis of cervical neoplasia has been prepared. Classic procedure involving biopsy may be replaced by in vivo non invasive optical biopsy. In this thesis, we are studying a model that estimates for each combination of 4 biological parameters, the spatiotemporal curves of diffused reflectance versus time, that are produced during the acetowhitening phenomenon. We aim to solve the reverse problem, estimating the bio-parameters given the curve, efficiently in terms of both accuracy and speed. Initially, we used several curve matching algorithms to decide which is optimal for the problem. Then, we used k-means clustering on the model dataset to reduce comparisons and lower procedure duration. Finally, we correlated curve features to bio-parameters and created a decision rule for instant bio-parameter estimation through optimization with decision regions. Judging from the estimation accuracy and the calculations on execution times, we have strong confidence in the future of the method.
Copyright © DIAS 2013
