Mol, Frank (2021) Shape Adaptive Kernel Density Estimation as smoothing method for MTObjects. Master's Thesis / Essay, Computing Science.
|
Text
mCS_2021_MolFN.pdf Download (5MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
Toestemming.pdf Restricted to Registered users only Download (119kB) |
Abstract
MTObjects is a segmentation method which segments astronomical images. The main difference between the state-of-the-art SExtractor and other techniques, is that MTObjects performs segmentation by means of creating a MaxTree of the image. MTObjects uses a smoothing technique, Gaussian Blur, in a significance test, to see whether a node in the MaxTree is significant. The σ parameter is set to 3. There lies an improvement in distinguishing noise from faint light-emitting objects and nested objects for MTObjects. Hence, adaptive smoothing methods could improve the workings of MTObjects. This thesis looks into different smoothing techniques to denoise astronomical images. Next to the Gaussian blur, we test the Perona Malik Diffusion method and develop a smoothing method based on Kernel Density Estimation. The smoother based on the Kernel Density Estimation can be made adaptive to the local intensity level and the local curvature. We investigate the performance of these techniques by comparing the smoothed noised image to a noiseless image, by means of denoising metrics. The denoising metrics used are the Peak-Signal-to-Noise-Ratio (PSNR), the Structural Similarity Index (SSIM) and the Normalized Root Means Squared Error (NRMSE). We perform a grid-search to investigate the optimal parameter settings. Both the Perona Malik Diffusion and the smoothing based on the Kernel Density Estimation outperform the Gaussian Blur.
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Supervisor name: | Wilkinson, M.H.F. and Biehl, M. |
| Degree programme: | Computing Science |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 13 Jul 2021 09:06 |
| Last Modified: | 13 Jul 2021 09:06 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/25195 |
Actions (login required)
 |
View Item |