Real Time NIR Imaging Image Enhancement by using 2D Frangi Filter via Segmentation

This paper presents the NIR imaging images enhancement by using 2D Frangi Filter
segmentation which specifically apply in biomedical NIR vein localization imaging. The unseen
subcutaneous vein causing clinical practitioner face the difficulties to perform intravenous
catheterization and thus lead to the needles tick injuries. There are few imaging techniques which
can be used for bein localization but the most widely used is Near Infrared (NIR) imaging due to
its non-invasive and non-ionizing properties. The input images from NIR imaging setup is
processed in order to enhance the vein visibility and contrast between vein and skin tissue. It is
required to filter noise from the display image using some image processing technique. This
work is done by applying image segmentation method to NIR venous image in order to extract
veins and eliminate the noise. First, the gray scale image was segmented to 10 pieces of fragment
plane with constant step size to produce 3 set of 2D planes. Second, these 3 sets of 2D planes
will then apply in Frangi filter in order to obtain the eigenvalue image structure. Lastly, a least
noise image is produce by this integrated plane through the 2D Frangi filter.