Patch-Based Adaptive Background Subtraction for Vascular Enhancement in X-Ray Cineangiograms

Area of Science:

• Medical Imaging
• Cardiovascular Imaging
• Image Processing

Background:

• Automatic vascular enhancement in X-ray cineangiography is vital for visualizing and quantifying coronary arteries.
• Accurate enhancement aids in diagnostic and interventional procedures for coronary artery diseases.

Purpose of the Study:

• To introduce a novel patch-based adaptive background subtraction method (PABSM) for automatic vascular enhancement in coronary X-ray cineangiography.
• To evaluate the effectiveness of PABSM in preserving vascular integrity and reducing noise.

Main Methods:

• PABSM utilizes vesselness and Gabor features to describe pixels.
• A classifier separates vascular and non-vascular regions, followed by dilation.
• Patch-based background synthesis fills removed vascular regions.

Main Results:

• The PABSM method achieved an area under the precision-recall curve of 0.7133 and a C-value of 0.9659.
• Performance was compared against six state-of-the-art vascular enhancement methods.

Conclusions:

• PABSM effectively enhances coronary arteries in cineangiograms, preserving topological structures in complex regions.
• The method reduces noise from non-uniform gray-level distribution and avoids motion artifacts.