| Title |
Piniginių monetų nudėvėjimo lygio nustatymas / |
| Translation of Title |
Level of wear estimation for currency coins. |
| Authors |
Gumbinas, Matas |
| Full Text |
|
| Pages |
67 |
| Keywords [eng] |
image processing ; illumination invariance ; multilevel threshold ; object segmentation ; coin wear |
| Abstract [eng] |
Estimating modern coin wear based on 2D digital images is a difficult problem. Lack of sufficiently large data sets, wide range of inter-class variance between the coin’s visual designs and unstable lightning or background conditions makes this problem increasingly complex to solve in real-life conditions. To help alleviate some of these issues, we introduce and evaluate new methods for coin background elimination, achieving illumination invariance and coin wear information extraction. Finally, we created and provided a new coin wear estimation dataset based on real-life coins images with various wear features. |
| Dissertation Institution |
Kauno technologijos universitetas. |
| Type |
Master thesis |
| Language |
Lithuanian |
| Publication date |
2020 |
