The target of palm vein recognition is not the texture of the skin surface (such as palm prints), but the vascular structure of the veins beneath the skin. This structure is genetically determined, remains stable throughout life after adulthood, and is unique to each person and each hand.

Currently, the DEWO palm vein module is compatible with both palm print and vein recognition, supporting the simultaneous reading of palm print and palm vein information.

The technical implementation of palm vein recognition can be roughly divided into four steps:

Step 1: Image Acquisition

1. Emitting Near-Infrared Light: The recognition device has a built-in near-infrared light emitter (typically with a wavelength of 700-1000 nanometers). This wavelength of light is safe for the human body and has good tissue penetration.

2. Selective Absorption: When near-infrared light shines on the palm: muscles, bones, and other tissues scatter or reflect some of the light. Deoxyhemoglobin in venous blood strongly absorbs specific wavelengths of near-infrared light.

3. Forming a Vein Image: Because veins absorb more light, the reflected light signals create a dark shadow at the location of the vein. The device's built-in infrared camera captures this image with the vein shadow.

Key Point: This is an "active imaging" process, relying on the inherent properties of blood, requiring no external staining or injection of any substances.

Step Two: Image Processing

The acquired raw image usually contains noise and needs contrast improvement.

Noise Reduction: Algorithms are used to filter noise from the image sensor and the environment.

Enhancement: The contrast between the vein shadow and the surrounding tissue is improved, making the vascular network clearer and more continuous.

Binarization: The image is processed into black and white, with black representing veins and white representing the background, resulting in a clear "skeleton" of the vein network.

Step Three: Feature Extraction – Converting the Image into Digital Code

The system does not store the entire vein image but extracts mathematical features that cannot be deduced from the original image. Commonly used features include:

Bifurcation and Endpoints of the Vein Network: Like intersections and ends in a road network, the number and location of these points are core features.

Blood vessel thickness and length: Dimensional information of different blood vessel segments.

Relative position and angle of blood vessels: Spatial geometric relationships between bifurcation points.

Texture features: Blood vessel distribution density and patterns in local areas.

These features are converted into a set of highly condensed, unique digital feature codes (templates). This template is typically only a few hundred bytes long and is irreversible (the vein image cannot be reconstructed from the template), greatly protecting biometric privacy.

Step four: Matching and Recognition – Comparing the "Password"

1. Registration (Entry): Upon first use, the system extracts the user's palm vein feature code and securely stores it as a template in a database or smart device terminal.

2. Verification/Recognition: When the user identifies again, the system collects and generates a feature code in real time.

Verification (1:1 comparison): Compares the feature code with the specific template corresponding to the user's claimed identity.

Recognition (1:N comparison): Searches for matches throughout the entire database.

3. Decision: The system calculates the similarity score between the real-time feature code and the stored template. If the score exceeds a preset threshold, authentication is successful; otherwise, it is rejected.

The DEWO palm vein module supports palm vein (palmprint) feature acquisition and comparison. 1:1 and 1:N (host computer comparison/server comparison) are supported. Our palm vein module has a false acceptance rate (FAR) of < 0.00001% and a false rejection rate (FRR) of < 0.01%.

Recognition speed reaches 150ms. Test data is based on 10,000 data points, with a device configuration of i3-7100U@2.4GHz.

Palm vein recognition angles: Hand rotation: horizontal rotation within 0~180°; Hand tilt: finger tilt within -20°~40°.

Advantages of palm vein recognition:

1. High anti-counterfeiting (liveness detection): Flowing blood (hemoglobin) is required to absorb near-infrared light and form an image. Forgery methods such as wax figures, photographs, and severed palms are completely ineffective. This is its core security advantage.

2. High Accuracy: Veins possess a complex internal three-dimensional structure with a vast amount of stable and unchanging feature information, resulting in an extremely low recognition error rate (e.g., 1 in 10,000).

3. Non-Contact: Due to its optical imaging nature, it eliminates the need for fingers or palms to press on the sensor, promoting hygiene and avoiding the impact of surface wear or residue on recognition.

4. High User Acceptance: The recognition process is natural, without psychological resistance, and is easily imperceptible, providing a user-friendly experience.

Palm vein recognition technology is essentially a combination of bio-optical imaging and pattern recognition. It cleverly utilizes the absorption characteristics of human blood to specific near-infrared light to create an externally invisible and difficult-to-replicate map of internal biometric features, which is then transformed into a secure and convenient "password" using advanced algorithms.

Hot-selling products