Jonge, Rick, G.D. de (2021) MALORQA: A Machine Learning approach to Objective Reflection Quality Assessment. Master's Thesis / Essay, Computing Science.
|
Text
Master_Thesis_v7.pdf Download (1MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
toestemming.pdf Restricted to Registered users only Download (123kB) |
Abstract
This thesis gives approaches to creating a machine learning model as an alternative to an analytical approach, which determine the smoothness of a surface visible in an image. It starts with describing what the analytical approach was set out to do, and how a machine learning model could improve on this. It then clarifies some important concepts, like what we mean by smoothness and how this can be visualized, and an existing algorithmic way of categorizing this smoothness. The creation of a dataset to identify this smoothness using machine learning is then explained, including the generation and rendering of the data. A first attempt at the machine learning program is defined, using machine learning on images of objects rendered using the same reflection lines as used in the Objective Rendering Quality Assessment (ORQA), of which the results are insufficient. A second attempt then builds on this by increasing the information in the input using normal mapping, and it is demonstrated that this improves both accuracy and speed of the learning process.
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Supervisor name: | Kosinka, J. and Hettinga, G.J. |
| Degree programme: | Computing Science |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 01 Apr 2021 09:38 |
| Last Modified: | 01 Apr 2021 09:38 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/24145 |
Actions (login required)
 |
View Item |