Related Experiment Videos
Edge detection in prostatic ultrasound images using integrated edge maps
R G Aarnink1, S D Pathak, J J de la Rosette
1Urology Biomedical Engineering Unit UIC/BME, University Hospital Nijmegen, The Netherlands. r.aarnink@uro.azn.nl
Ultrasonics
|July 4, 1998
Summary
This study introduces a new algorithm for edge detection in prostate ultrasound images. By using multiple scales of resolution, it improves the accuracy of locating anatomical boundaries.
Area of Science:
- Medical Imaging
- Image Processing
- Ultrasound Technology
Background:
- Previously developed a non-analytical operator for prostate contour determination using minimum and maximum filters.
- Implemented a technique to enhance boundary determination in prostatic ultrasound images by adjusting edge detection parameters.
Purpose of the Study:
- Investigate an algorithm for detecting grey level transitions at multiple resolutions.
- Improve edge detection and localization in prostate ultrasound images.
Main Methods:
- Investigated prefilter settings and edge detection parameters in test and real ultrasound images.
- Utilized local standard deviation to identify homogeneous regions for coarse filtering.
- Applied smaller filter sizes to areas with higher deviation to preserve grey level transitions and improve edge localization.
Main Results:
- Analysis showed that increasing filter sizes merge image areas and obscure less pronounced edges.
- Two scales of resolution improved edge localization, particularly in areas with increased local standard deviation when using smaller filters.
Conclusions:
- Proposed an edge detection method as a valuable pre-processing step for medical image interpretation.
- Demonstrated that adapting input parameters to signal information enables application across different imaging modalities.
- Discussed limitations, leading to a combined localization and delineation algorithm for contour refinement.