Palm vein recognition uses near-infrared (IR) light to capture the unique vein pattern beneath the skin of the palm. These internal patterns are invisible to the naked eye, making them highly secure and extremely difficult to replicate.

X-Telcom’s ( BioWavePass) Palm Vein Biometric Technology goes beyond traditional IR-only systems. It captures IR vein patterns and RGB surface features at the same time. By combining internal vein data with external palm features, the system builds richer biometric data, improving recognition accuracy, speed, and resistance to spoofing in large-scale databases.

There is no strict size limitation. The scanning distance must be between 5–15 cm. In this range, palms with width 5–10 cm and length 15–22 cm can be stably captured, covering roughly children aged 6–12 through adults with extra-large hands.

The algorithm includes liveness and quality assessment modules. It scores occlusion rate, image contrast, vein clarity, and other parameters. If the image does not meet liveness or quality thresholds, recognition is automatically rejected.

Palm vein and palm print patterns are considered basically stable after age 6. Technically there is no hard lower limit. However, in many payment scenarios, enrollment is restricted to users aged 14 or above to match real-name account regulations. Users over 75 can still use the system if their palm veins are not severely degraded.

The system uses 850 nm near-infrared LEDs and high-sensitivity global-shutter CIS sensors, penetrating 2–3 mm under the skin. For reduced blood flow scenarios, exposure time can be extended by 10–20 ms and sensor gain increased for compensation.

RGB and IR resolution is 720 × 1024.

Yes. Users need to place their palm above the module. The system uses distance sensing and palm ROI framing to judge palm position and provide UI guidance on screen.

Multi-frame fusion is used to compensate for natural motion. The system can tolerate hand movement speeds up to 0.5 m/s during scanning.

Temperature: −10°C to 50°C

Humidity: 0–95%

Illumination: 0–50,000 Lux (module must not be exposed to direct sunlight)

Under normal hardware and network conditions, a single module completes each recognition in 1–1.5 seconds, including image capture, liveness and quality filtering, feature extraction, upload, and result return.

Overall recognition accuracy is 99.5%.

RGB palm print: FRR 1.002% at FAR 1e-6

NIR palm vein: FRR 1.52% at FAR 1e-6

Ours is RGB + NIR combined: Probability of both failing simultaneously is about 1e-11.

Standby: 1.5 W

Instant during scanning: up to 7.5 W

Average: about 2.5 W

Data is encrypted at the device using AES-256 with license signature. Transmission uses SSL/TLS encryption.

The system mainly provides algorithm matching capability. Algorithm services store palm feature vectors and palm indexes. Other data is managed by the customer based on application needs. Customers are advised to store original RGB and IR palm images collected during enrollment according to local regulations, on their own servers. X-Telcom does not access or manage customer biometric data.

Yes, HTTPS with SSL encryption is used between devices and backend servers.

Palm feature information is encrypted using AES-256-CBC for both palm vein and palm print data.

Each record is slightly over 4 KB, including about 2 KB for palm vein and 2 KB for palm print. One million records require roughly 4 GB of storage. If image caching is used, 1 million records can require up to 1 TB.

Network bandwidth

Server CPU/GPU performance

Client device processing capability

Recognition is usually completed within 1 second, including client-side processing and server-side matching.

At least 100 Mbps bandwidth is recommended for stable performance.

Tests include palm photos, screen replays, 3D-printed hands, gloves, and animal skin. Liveness detection is based on combined RGB and IR imaging.

Free small-model algorithm for up to 10,000 IDs.

Paid large-model licenses for 100,000 IDs and 1,000,000 IDs.

On our current domestic private servers, we already have over 5 million user IDs, and our WeChat collaboration has reached over 300 million users. Overseas customers have also purchased licenses for over 20 million users.

In the large-scale comparison model, the algorithm itself enforces a strict limit on the total number of palm records that can be registered. Even though X-Telcom cannot see or access the customer’s database content, the algorithm will automatically restrict registrations based on the licensed capacity. For example, if a 100,000-user license is purchased, the system can register a maximum of 100,000 users in their own databse.

For the free small-model version with a 10,000 ID license, there is no hard system cap on the number of registrations. However, if the number of registered users exceeds about 20,000, the comparison error rate will increase significantly. The only way to achieve around 99.5% recognition accuracy is to stay within 10,000 registered users. No technical support or service will be provided for error rates caused by exceeding the specified ID limit.

Yes. After purchase, the license is permanently valid with no expiration date. It is a one-time payment, not a subscription or recurring fee.

The ID quantity refers to the number of registered users in the system. It is only related to how many users are enrolled, not to how many transactions are made.

For example, if you register 100,000 users, then your license ID quantity is 100,000. This number has nothing to do with daily or monthly transaction volume.

If a registered user is deleted from the system, that ID slot becomes available again and can be used to register a new user.