Improved automatic exposure control using morphology-based disturbance recognition

R.I. Gaasbeek, R.J.R. van der Maas, M. den Hartog, A.G. de Jager

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

1 Citation (Scopus)
2 Downloads (Pure)

Abstract

In medical X-ray imaging, the detector intensity heavily influences the signal-to-noise ratio, and thus the image quality. Consequently, image quality and patient dose are dependent on the performance of the Automatic Exposure Control. Introducing large opaque objects to the image, which can be considered disturbances for the dose control, leads to a loss of image quality (overexposed tissue) as well as an increased patient dose. The effect of scatter-radiation makes it difficult to exclude these disturbances from the image using simple thresholding. In this work, a morphology-based filter is proposed as a pre-processing step for the Automatic Exposure Control leading to a superior disturbance exclusion. The algorithm has been verified in a real-time environment and it is shown to be robust against large disturbances in the X-ray images.
Original languageEnglish
Title of host publicationProceedings of the 2014 IEEE International Symposium on Biomedical Imaging (ISBI)
Place of PublicationPiscataway
PublisherInstitute of Electrical and Electronics Engineers
Pages1271-1274
Number of pages4
ISBN (Electronic)978-1-4673-1961-4
DOIs
Publication statusPublished - 2014
Event11th IEEE International Symposium on Biomedical Imaging (ISBI 2014) - Renaissance Beijing Capital Hotel, Beijing, China
Duration: 29 Apr 20142 May 2014
Conference number: 11
http://biomedicalimaging.org/2014/

Conference

Conference11th IEEE International Symposium on Biomedical Imaging (ISBI 2014)
Abbreviated titleISBI 2014
Country/TerritoryChina
CityBeijing
Period29/04/142/05/14
Internet address

Fingerprint

Dive into the research topics of 'Improved automatic exposure control using morphology-based disturbance recognition'. Together they form a unique fingerprint.

Cite this