Hessian filter-assisted full diameter at half maximum (FDHM) segmentation and quantification method for...

Area of Science:

• Biomedical Imaging
• Optical-Resolution Photoacoustic Microscopy
• Quantitative Vascular Analysis

Background:

• Optical-resolution photoacoustic microscopy is ideal for angiographic studies.
• Accurate vessel segmentation is crucial for quantitative vascular analysis.
• Existing Hessian filter methods show variable accuracy due to multi-kernel strategies.

Purpose of the Study:

• To develop an improved vessel segmentation algorithm for photoacoustic microscopy.
• To enhance the accuracy and consistency of quantitative vascular analysis.
• To enable more sensitive detection of subtle vascular changes.

Main Methods:

• Developed a Hessian filter-assisted, adaptive thresholding vessel segmentation algorithm.
• Validated performance using a digital phantom and in vivo images.
• Tested subtle vessel change detection in longitudinal studies with blood pressure agents.

Main Results:

• The proposed algorithm achieved a consistent accuracy of 0.987, independent of kernel selection.
• Detected twice the vasoconstriction compared to the Hessian filter method in antihypotensive studies.
• Identified a 21.2% vasodilation in antihypertensive studies, where the Hessian filter method failed.

Conclusions:

• The developed algorithm offers superior and consistent accuracy for vessel segmentation.
• It significantly improves the detection of subtle vascular changes compared to traditional methods.
• This advancement holds potential for pushing the limits of quantitative imaging in angiographic applications.