Class-based Nonlinear Enhancement Strategy and Its Validation: An Application for Digital Subtraction Angiography (DSA)

L. Wei, D. Kumar, J. Coleman, R. Turlapthi, and J. Suri (USA)


DSA, image enhancement, nonlinear normalization, SNR, observer


Digital Subtraction Angiography (DSA) is a well established modality for the visualization of blood vessels in the human body. DSA helps interventional clinicians make decisions in analyzing vascular diseases. Though, there is a clear motivation to improve the image quality for these procedures, it suffers from challenges like system X-ray noise and motion artifacts due to patient movement. This paper presents a class-based image enhancement technique so-called nonlinear normalization to enhance the blood vessel there by suppressing the background. A lookup table (LUT) is developed in real time which enhances the dye injected blood vessels retaining its translucency and suppressing the background. Our protocol is evaluated on a database of 73 subjects by two different strategies: (a) using three trained human observers and (b) quantitatively, using signal-to noise (SNR) measurement. Using our performance evaluation protocol, SNR of the DSA embedded with nonlinear enhancement method demonstrates an improvement of 24.87% over conventional DSA. We validate our algorithm using GE’s PMMA phantoms. Our system runs on Eigen’s DSA workstation using C++ in Windows environment.

Important Links:

Go Back