The venous system plays a critical role in the overall circulatory network by facilitating the constriction and dilation of blood vessels, storing large volumes of blood for redistribution, and even regulating cardiac output. However, varicose veins (VV) pose a significant health concern, characterized by abnormally enlarged, twisted, and darkly colored veins primarily affecting the lower limbs [1]. VV occurs as a result of diminished flexibility and weakened valves within the veins, which leads to improper blood flow and subsequent vein dilation. This condition often causes discomfort, pain, skin changes, and the development of ulcers [2], [3], [4].

Hyperspectral imaging (HSI) is an advanced non-invasive optical imaging technique that has the potential to revolutionize medical imaging research and clinical applications. It operates by collecting and analyzing spectral information across the electromagnetic spectrum [5,6]. While HSI typically covers visible wavelengths between 450 and 700 nm, it can also extend into the infrared (>700 nm) and ultraviolet (<450 nm) ranges [7,8]. The advantage of HSI over conventional imaging lies in its ability to provide spectral reflection or absorption characteristics of the imaged object, which are represented as spectral channels within a hypercube of image data.

Previous studies have successfully utilized HSI for various medical applications the disease studied (prostate [9], ovaries [10], breast [11], tongue cancer [12], skin and lung cancer [13] or oral cancer [14]), and the applied processing methods to analyze the HS data (as in the application to larynx cancer [15]). Hyperspectral brain cancer imaging classification PLOS, including pathological slide imaging, liver thermal monitoring following radiofrequency ablation, breast cancer detection, and the determination of oxygen levels in blood ex vivo. These studies demonstrated promising sensitivity and specificity in detecting and characterizing these conditions using HSI [16], [17], [18], [19], [20], [21], [22].

The objective of this study is to evaluate the performance of a narrow band filter hyperspectral imaging system to specifically identify varicose veins at 530 nm and the main veins of the lower limbs at 780 nm. By employing quantitative phase analysis and a k-means clustering algorithm, we hypothesize that image processing techniques can accurately differentiate varicose veins from normal veins in an automated fashion. This approach holds the potential for both semi-automated detection using k-means clustering and fully automated detection using quantitative phase clustering [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34].

In summary, this research aims to leverage the capabilities of hyperspectral imaging combined with image processing algorithms to effectively identify and delineate varicose veins in the lower limbs. The potential benefits include improved diagnostic accuracy, treatment planning, and overall patient care in the field of vascular medicine.