K-space Compounding for Improved Endocardial Border Detection
- Resource Type
- Authors
- Gregg E. Trahey; Melissa LeFevre; Anna Lisa Crowley; Jayne Cleve; Nick Bottenus
- Source
- 2019 IEEE International Ultrasonics Symposium (IUS).
- Subject
- Channel (digital image)
business.industry
Computer science
k-space
Pattern recognition
030204 cardiovascular system & hematology
Endocardial border
01 natural sciences
Visualization
03 medical and health sciences
Speckle pattern
0302 clinical medicine
Compounding
0103 physical sciences
Medical imaging
Artificial intelligence
business
010301 acoustics
Image resolution
- Language
Visualization of the endocardial border is essential in determining left ventricular and overall cardiac function. Ultrasound imaging systems typically rely on image post-processing for contrast and detail enhancement. Spatial compounding is a physics-based processing method in the channel domain that exchanges resolution for texture reduction. We propose k-space compounding, a variation on the conventional technique enabled by recovery of the complete data set that allows for more aggressive spatial compounding.We performed cardiac scanning on 25 volunteers and patients at the Duke University Hospital to evaluate the degree of compounding useful for diagnostic imaging. Of these, 18 subjects were included for both qualitative and quantitative analysis. We found that compounding improved endocardial border detection according to the generalized contrast-to-noise ratio in all cases and more aggressive compounding improved further in 10 out of 18 cases. Three expert reviewers evaluated the images for their usefulness in several diagnostic tasks and determined that some spatial compounding is usually preferred, but more or less compounding was sometimes beneficial. K-space compounding provides a wide continuum of processing options along the resolution and texture trade-off to enable a clinician to optimize cardiac images.