Main technical features of vein identification
Author: huifan Time: 2022-12-22
Vein recognition is an emerging infrared biometric technology, which is based on the characteristic of deoxygenated heme in venous blood absorbing near-infrared rays or far-infrared rays radiated by human body, taking the vein distribution map of the back of hand (or finger back, finger belly, palm, wrist) with infrared camera of corresponding wavelength range, filtering and enhancing through normalization, denoising and other pre-processing, dividing and refining and repairing the vein pattern, then extracting its characteristics, and then The features are then extracted and matched with the pre-registered database or the features stored on the IC card to determine the identity of the individual. It uniquely identifies a person because each person's vein distribution has the uniqueness similar to fingerprints and persists after adulthood. In addition, it has advantages that other biometric identification technologies do not have, and thus has a wide range of application prospects and has gained the attention of scholars.
I. Introduction of principle
Vein recognition mainly uses the structure of veins to carry out identity recognition. Since the vein vein contains a large amount of characteristic information, it can be used as the object of verification. The principle of palm vein identification is also the same as that of palm vein identification by using the different absorption characteristics of infrared light of specific wavelength between vein vessels and muscles and bone bones to perform vein angiography. Due to the thick palm of the hand, the infrared light is usually not transmitted, so the only method used is reflection angiography. The infrared light is shone on the back of the hand, the areas with veins absorb the infrared light and reflect it dimly, while the muscular and bony areas reflect it strongly, thus creating a picture of the veins. The vein pattern is difficult to be faked inside the body.
(A) Live identification
When using the dorsal hand vein for identification, the image feature of the dorsal hand vein is acquired, which is a feature that exists only when the dorsal hand is alive. In this system, the vein image feature is not available for the non-living back of the hand and thus cannot be identified and thus cannot be forged.
(ii) Internal features
When the dorsal hand vein is used for authentication, the vein image features are obtained from the inside of the back of the hand and not from the surface of the hand. Therefore, there are no barriers to identification due to damage, wear and tear, dryness or too wetness of the back of the hand surface.
The back of the hand does not need to contact with the device, and recognition can be completed with a gentle release. This way does not have the problem of unhygienic when the hand touches the equipment and the security problem brought by the finger surface features that may be copied, and the well avoids the psychological discomfort of being treated as the object of examination, and at the same time does not fail to identify after contamination by dirty things. The palm vein method is suitable for almost all users because the veins are located inside the palm of the hand and the degree of influence of external factors such as temperature is negligible. Good user acceptance. In addition to not requiring direct contact with the scanner surface, this non-invasive scanning process is both simple and natural, alleviating any possible user resistance due to concerns about hygiene or usage hassles.
(iv) High security level
Because of the previous 3 aspects of live identification, internal characteristics and non-contact characteristics, it ensures that the user's dorsal hand vein characteristics are difficult to be forged. So the dorsal hand vein recognition system has high security grade and is especially suitable for use in places with high security requirements.
Third, vein recognition algorithm
It mainly includes 3 major parts: vein image acquisition; vein image preprocessing and vein recognition. The image preprocessing part mainly consists of Gaussian low-pass filtering, Gaussian high-pass filtering, Gap value processing, bilinear filtering and improved median filtering. By experimenting with 5000 samples, the recognition rate can reach 94.88%.