3D Finger Vein Biometric Recognition Protocol: It compares and matches the extracted 3D Finger Vein patterns against a pre-registered pattern database to recognize the individual. Generally, it involves three steps: normalization, feature extraction, and classification.
Once the ROI is extracted, images are enhanced by removing noise and/or lightening the vessels, a technique defined by the Contrast Limited Adaptive Histogram image that will be extracted, stored as a feature vector and processed for recognition. These features contain only relevant information such as orientation, length, and location. For instance, the vein pattern can be detected from the original image, reduced to 1 pixel-width (thinned) to form a skeleton and stored as a biometric feature. The vein recognition method will be the Minutiae Points Technique and for the skeleton extraction, we will use the multi-scale Gaussian Matched Filter.