Visualization of palm veins in NIR images, application in clinical injection and identification

Finding appropriate location of veins is important in two areas of clinical injection and identification. This is non-invasive method that makes it applicable in these areas especially in emergency injections. In identification, pattern of veins could be used. Images from external limbs are captured, so in this paper NIR images from the back of hand are used. These images are processed and segmented with two methods, maximum curvature and level set. Maximum curvature algorithm has good accuracy for injection after approval of doctors but the results from level set method is not acceptable. In identification, correlation used to check validity of methods. Precision of two algorithms is 100% if the position of the hand is fixed. Maximum curvature and level set methods tolerate rotation of about 12° and 6° in their worst case respectively as there may be unpredictable displacement in hand position, which happens because of differences in structures of algorithms.