Locating abnormalities in brain blood vessels using parallel computing architecture.

Area of Science:

• Medical Imaging and Visualization
• Computer-Aided Diagnosis
• Neuroscience

Background:

• Computed tomography (CT) and magnetic resonance imaging (MRI) provide 2-D images, complicating the interpretation of 3-D anatomical structures and abnormalities.
• Accurate 3-D reconstruction of medical images is computationally intensive, time-consuming, and costly.
• Detecting intricate internal features like blood vessels in 3-D remains a challenge in computer-aided diagnosis.

Purpose of the Study:

• To develop an efficient volume visualization framework for 3-D medical image reconstruction.
• To improve the speed and reduce the cost of 3-D visualization for diagnostic purposes.
• To enhance the detection of internal anatomical features, specifically blood vessels, for improved medical diagnosis.

Main Methods:

• Implementation of a volume visualization framework utilizing Compute Unified Device Architecture (CUDA) for accelerated processing.
• Augmentation of the ray casting technique to enhance image quality and visualization speed.
• Development within the Microsoft .NET environment for seamless interoperability with other medical tools.

Main Results:

• The framework successfully performed immediate 3-D reconstruction of brain datasets.
• Obvious mappings of internal brain features, including blood vessels, were achieved.
• Evaluation on 109 MRA datasets demonstrated reliable and instantaneous identification of potential blood vessel blockages.

Conclusions:

• The proposed CUDA-based framework offers a cost-effective and rapid solution for 3-D medical image visualization.
• It significantly improves the ability to detect internal anatomical structures and pathologies, such as cerebral vascular blockages.
• The framework's interoperability and efficiency make it a valuable tool for computer-aided diagnosis in neuroimaging.