Change search
ReferencesLink to record
Permanent link

Direct link
Automatic X-Ray Parameter Analysis and Reduction
KTH, School of Electrical Engineering (EES), Sound and Image Processing.
2011 (English)Student paper other, 20 credits / 30 HE creditsStudent thesis
Abstract [en]

This thesis describes research in the area of parameter analysis for X-ray imaging. This work was performed at Philips Healthcare in Best (the Netherlands) as a nal project for the Master study at the Sound and Image Processing Laboratory at Kungliga Tekniska Hogskolan (KTH), Stockholm. The objective of this project is to provide methods for automatic parameter analysis and reduction for X-ray tuning. These methods can be used to reduce the amount of parameters involving in X-ray tuning. X-ray processing is performed via a black-box process and parameter analysis consists in looking at the impact on the resulting X-ray image. The visual quality of this image depends on parameter tuning. With a large number of parameters, analysing their visual impact directly is not feasible, which is why objective image quality (OIQ) assessment is used to get numerical results. Several image quality assessment models are reviewed leading to further research in the full-reference and no-reference model approaches. Both assessments are explored with investigation of four dierent full-reference metrics, namely the Peak-Signalto-Noise Ratio (PSNR), the Structural Similarity (SSIM), the Visual Information Fidelity (VIF) and one using wavelet properties that we have called Wavelet Method (WM), and three no-reference metrics: noise, contrast and sharpness. Search algorithms are used to get a set of parameters which give the same image quality (using OIQ) such that dimensionality reduction can be performed. Several search algorithms are reviewed from the simplest (looking at the function evaluations of all points) to the most sophisticated algorithms for global optimization (e.g. genetic algorithm). Depending on the function to optimize, dierent algorithms are used. Finding corrolated parameters, or parameters that have no impact on the image quality is the way to reduce the amount of parameters. Principal Composent Analysis (PCA) which is one of the most common method for dimensionality reduction is performed on the results of combination of OIQ and search algorithms. For each step of the project, we test the assessments or the algorithms on some examples to validate the used procedure. We will nally test all our methods with one IP function which acts like a real X-ray process. The results will enable us to see if parameter analysis and reduction is feasible.

Place, publisher, year, edition, pages
2011. , 54 p.
EES Examensarbete / Master Thesis, XR-EE-SIP 2011:008
National Category
Engineering and Technology
URN: urn:nbn:se:kth:diva-55378OAI: diva2:471540
Available from: 2012-01-19 Created: 2012-01-02 Last updated: 2012-03-16Bibliographically approved

Open Access in DiVA

fulltext(7922 kB)161 downloads
File information
File name FULLTEXT01.pdfFile size 7922 kBChecksum SHA-512
Type fulltextMimetype application/pdf

By organisation
Sound and Image Processing
Engineering and Technology

Search outside of DiVA

GoogleGoogle Scholar
Total: 161 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 162 hits
ReferencesLink to record
Permanent link

Direct link