Abstract P226: Vessel Segmentation Based on Multi-Thresholding for Diagnostic Analysis of Microcirculation
Introduction: The microcirculation has become a key target for study and assessment of tissue perfusion and oxygenation. Detection and assessment of the microvasculature (arterioles, capillaries, venules) using videomicroscopy from the oral mucosa can provided a measurement of microvascular density in each single frame. Information pertaining to the density of these microvessels within a field of view can be used to quantitatively monitor and assess the changes occurring in tissue oxygenation and perfusion over time. This information, however, is extremely difficult to rapidly quantitate in near real-time. If automated, it could be used for real-time diagnostic and therapeutic assessment of resuscitation.
Methods: We designed an automated image processing method of microcirculatory video recordings to segment, estimate the density, and identify the distribution of blood flow in microvessels. The algorithm consists of three main steps: pre-processing, image segmentation and post-processing. In the pre-processing step, image enhancement is performed through acquiring adaptive histograms of the image. For sharpening edges and denoising, speckle reducing anisotropic diffusion filtering is used. The main technique used in the segmentation step is multi-thresholding and pixel verification based on geometric and contrast parameters. In the post-processing phase, controlled region growing is applied to fill vessel discontinuities.
Results: The algorithm was applied to a data bank of videos from healthy and critically ill humans/animals obtained using side stream darkfield imaging (Microvision Medical, Inc). Segmentation results were compared and validated using a blinded detailed inspection by experts who used a commercial image analysis software program (Microvision Medical Inc), which requires off-line video analysis. The experimental algorithm was found to extract approximately 97% of microvessels in every frame compared to the current software.
Conclusion: The proposed method is an entirely automated process that can perform pre-processing, segmentation, and microvessel identification without human intervention. The method may allow for assessment of microcirculatory abnormalities occurring in critically ill patients.