Palm vein recognition is becoming increasingly relevant for e-KYC, digital identity, banking, biometric authentication and palm payment.
But how does a palm vein recognition system actually work?
Unlike conventional fingerprint recognition, palm vein technology uses near-infrared (NIR) imaging to capture vascular characteristics beneath the skin. X-Telcom’s palm recognition technology goes further by combining NIR palm vein information with RGB palmprint characteristics, creating a multimodal approach to biometric authentication.
From capturing the palm to identifying a user within a large database, several technologies work together.
Palm Capture → Detection → Liveness Detection → Feature Extraction → Biometric Matching → Identity Result
Let’s look at each stage.
What Is Palm Vein Recognition?
Palm vein recognition is a biometric technology that uses the vascular characteristics of a person’s palm for identity verification or identification.
When a palm is presented above a compatible scanner, near-infrared light is used as part of the imaging process to capture information associated with the vein patterns beneath the skin.
These characteristics can then be processed by a biometric algorithm and converted into features used for matching.
Unlike a password, card or QR code, the biometric credential is associated with the person themselves.
This makes palm recognition particularly interesting for:
- e-KYC
- Digital identity
- Palm payment
- Banking authentication
- Healthcare identification
- Government identity
- Access control
- Self-service applications
Step 1: Capturing the Palm
The first stage is obtaining a high-quality image of the user’s palm.
X-Telcom palm recognition technology uses both:
- RGB imaging for palmprint characteristics
- NIR imaging for palm vein characteristics
The user presents their palm within the recognition area without needing to physically touch the scanner.
This contactless interaction is one of the major differences between palm recognition and conventional fingerprint readers.
For products such as the X-Telcom PalmVein01, the recommended palm recognition distance is approximately 5 to 15 cm, allowing users to authenticate with a simple hand presentation.
Step 2: Palm Detection
Before biometric matching can begin, the system first needs to detect and locate the palm correctly.
The detection algorithm identifies the palm within the captured image and prepares the relevant area for subsequent biometric processing.
This stage is important because users will naturally present their hands at slightly different positions and angles.
X-Telcom’s large-scale palm detection testing achieved:
99.64% Precision @ 99.96% Recall
Reliable palm detection helps create a fast and natural contactless user experience.
Step 3: Liveness Detection
Finding a palm in an image does not automatically mean that a genuine person is presenting it.
For higher-security applications, the system also needs liveness detection, sometimes referred to as Presentation Attack Detection (PAD).
The purpose is to distinguish genuine palm presentations from attempted spoofing.
X-Telcom’s palm liveness testing has included tens of millions of presentation attack samples, including:
- Printed paper fake hands
- Plastic fake hands
- Silicone fake hands
- Gloves
- Other spoof samples
The documented test performance reached:
98.69% TAR @ FAR 1×10⁻³
This layer becomes particularly important when palm recognition is used for banking, payment, e-KYC and other high-security identity applications.
Step 4: Extracting Palmprint and Palm Vein Features
Once the palm has been successfully captured and processed, the system extracts biometric features.
X-Telcom uses two biometric modalities:
RGB Palmprint
RGB imaging captures characteristics associated with the visible palm.
NIR Palm Vein
Near-infrared imaging captures information associated with vascular characteristics beneath the skin.
The resulting biometric features are converted into digital data that can be used by the matching algorithm.
For X-Telcom’s current architecture, the biometric feature data is approximately 4 KB, comprising roughly:
- 2 KB IR feature data
- 2 KB RGB feature data
The system therefore does not need to compare conventional photographs pixel by pixel during every authentication attempt. Instead, the algorithm works with extracted biometric features designed for matching.
Step 5: Biometric Matching
After feature extraction, the system compares the newly captured biometric information with previously enrolled biometric data.
There are two important types of biometric matching.
1:1 Biometric Verification
In 1:1 verification, the system already has a claimed identity.
It answers the question:
"Is this palm the same person as the identity being presented?"
This approach can be used where another identifier is already available.
1:N Biometric Identification
In 1:N identification, the user does not necessarily need to provide another identifier first.
The system searches the biometric database to answer:
"Whose palm is this?"
This capability is particularly important for applications such as palm payment and large-scale biometric identity systems.
A user can present their palm, allowing the biometric system to identify the corresponding registered identity within the database.
How Accurate Is Palm Vein Matching?
Biometric accuracy should not simply be described as "99% accurate."
Professional biometric systems are evaluated using metrics including:
- FAR: False Acceptance Rate
- FRR: False Rejection Rate
- TAR: True Acceptance Rate
X-Telcom’s large-scale palm matching testing achieved:
97.8779% TAR @ FAR 1×10⁻⁸
A FAR of 1×10⁻⁸ corresponds to approximately one false acceptance per 100 million comparisons under the stated test conditions.
The palm matching evaluation used data at the scale of hundreds of millions of records, making large-scale matching performance an important part of the evaluation.
Why Combine Palm Vein and Palmprint Recognition?
X-Telcom does not rely on only one source of palm information.
The technology combines:
RGB Palmprint + NIR Palm Vein
This multimodal approach provides different biometric information from the same palm presentation.
For X-Telcom’s small-model algorithm, the current default threshold performance includes:
| Biometric Modality | Performance |
|---|---|
| RGB Palmprint | FRR 1.002% @ FAR 1×10⁻⁶ |
| NIR Palm Vein | FRR 1.52% @ FAR 1×10⁻⁶ |
Using both modalities provides the biometric matching system with additional information when making an authentication decision.
This is one of the important technical differences between a professional palm recognition solution and a basic camera-based hand recognition system.
How Does Palm Vein Enrolment Work?
Before a user can be recognised, they first need to be enrolled.
A simplified enrolment process can look like this:
Present Palm → Capture RGB + NIR → Quality Check → Extract Features → Create Biometric Record → Link to User ID
Once enrolled, the user can return later and present their palm again.
The new biometric features are extracted and compared with the registered biometric data.
Depending on the application, the resulting identity can then be connected to:
- A customer account
- A payment account
- A digital identity
- An employee record
- A healthcare record
- An access permission
The palm recognition system therefore acts as the biometric identity layer within a wider application.
How Does Palm Vein Recognition Scale Beyond 10,000 Users?
Small biometric projects and large commercial deployments require different matching architectures.
X-Telcom provides two algorithm models.
Small Model
The small model is designed primarily for samples, MVPs, proof-of-concept projects and smaller applications.
It supports up to:
10,000 registered user IDs
This allows customers to develop and test their application before investing in a large-scale algorithm environment.
Large Model
For projects requiring more than 10,000 registered users, X-Telcom provides a commercial large-model architecture.
It is designed for applications ranging from hundreds of thousands to millions of registered identities.
The large model uses additional algorithm models and matching scores and can be deployed on a private server infrastructure.
Why Does QPS Matter for Large-Scale Palm Recognition?
For large biometric databases, capacity is only one consideration.
Another important factor is QPS, or Queries Per Second.
QPS represents the number of biometric matching requests the infrastructure is designed to process per second.
For example, a small e-KYC deployment and a national payment platform may have completely different peak transaction requirements, even if both use the same biometric technology.
X-Telcom reference configurations therefore scale according to both registered user capacity and required QPS.
| Database Capacity | 100K | 500K | 1M |
|---|---|---|---|
| QPS | 10 | 50 | 100 |
| RAM | 4 × 32 GB | 8 × 32 GB | 8 × 32 GB |
| SSD | 1 TB | 2 TB | 2 TB |
| GPU | NVIDIA L4 | NVIDIA L4 | 2 × NVIDIA L4 |
The final server architecture should be determined according to the specific project’s user capacity and transaction requirements.
Can Palm Vein Recognition Be Privately Deployed?
Yes.
For large-model applications, X-Telcom supports private server deployment.
This can be particularly important for:
- Banks
- Payment providers
- Government organisations
- Healthcare networks
- Large enterprises
- National identity projects
Private deployment allows the customer to operate the biometric matching environment within its own infrastructure or designated server environment.
The biometric layer can then integrate with the customer’s application through the appropriate SDK or API architecture.
How Does Palm Vein Recognition Work for e-KYC?
Palm vein recognition can become part of a wider digital identity or e-KYC workflow.
For example:
Customer Identity → Palm Enrolment → Liveness Detection → Biometric Record → Account Link → Future Palm Authentication
Palm recognition does not replace the entire KYC process.
Instead, it can provide a strong biometric mechanism for identifying and re-authenticating an already enrolled customer.
This is particularly useful where customers repeatedly interact with physical locations such as banks, healthcare facilities, government services or self-service terminals.
How Does Palm Vein Recognition Work for Palm Payment?
Palm payment takes the same biometric identification principle and connects it to a payment ecosystem.
A simplified transaction could look like:
Present Palm → Liveness Detection → 1:N Identification → Identify Payment Account → Authorise → Process Payment
However, the biometric system is only one part of the complete solution.
A commercial palm payment ecosystem typically requires:
Palm Recognition + Digital Identity + Payment Account + Payment Gateway / Acquirer + Certified Payment Hardware
This means successful palm payment deployment normally requires cooperation between biometric technology providers, banks, fintech companies, acquirers, payment gateways and payment processors.
X-Telcom focuses on providing the palm recognition hardware, algorithms, SDK/API and biometric integration technology required within this ecosystem.
Frequently Asked Questions About Palm Vein Recognition
Does palm vein recognition require physical contact?
No. X-Telcom palm recognition devices are designed to capture palm biometric information without requiring the user to touch the scanner.
What is the difference between palmprint and palm vein recognition?
Palmprint recognition uses visible characteristics of the palm captured through RGB imaging. Palm vein recognition uses NIR imaging to capture information associated with vascular characteristics beneath the skin.
X-Telcom combines both modalities.
Does X-Telcom support Linux and Android?
Yes. X-Telcom provides SDK support for multiple platforms, including Android, Windows and Linux, depending on the product and project requirements.
Can palm vein recognition perform 1:N identification?
Yes. X-Telcom’s large-model architecture is designed to support large-scale biometric identification where a palm can be matched against a registered user database.
How many users can the free testing model support?
The small-model environment supports up to 10,000 registered user IDs, making it suitable for samples, MVP development and proof-of-concept testing.
Can the algorithm be deployed on the customer’s own server?
Yes. X-Telcom supports private server deployment for large-model commercial projects.
From Palm Capture to Digital Identity
Palm vein recognition is much more than simply taking an infrared image of a hand.
A complete biometric authentication process combines:
RGB + NIR Capture → Palm Detection → Liveness Detection → Feature Extraction → Biometric Matching → Identity Authentication
For large-scale applications, this must then be supported by appropriate algorithm architecture, database capacity, QPS performance, server infrastructure and application integration.
This is what enables palm recognition to move beyond conventional access control and into applications such as e-KYC, digital identity, banking and palm payment.
X-Telcom provides palm recognition hardware, biometric algorithms, SDK/API integration and scalable matching technology for organisations developing the next generation of biometric identity solutions.
