In 2024, Americans lost a total of $47 billion to identity fraud and scams. This included $27 billion from traditional identity fraud, which affected over 18 million people in the US. The remaining $20 billion was lost to scams involving social engineering tactics. As fraud threats grow and customers expect seamless remote onboarding, organizations are turning to industry leaders like Mitek Systems, known for its expertise in liveness detection and document fraud prevention. One of the most effective tools in this effort is document liveness detection — a fast and automated way to confirm that an ID is physically present during the verification capture.
What is document liveness detection?
Document liveness detection is the process used to confirm whether an image of an identity document like a driver’s license, passport, or other official form of identification, was captured from a real, physical ID. It helps detect whether the document was instead spoofed using a printout, an image displayed on a screen, or a tampered version of a document.
Types of document fraud
Screen replay attacks are one of the most common forms of document fraud. Replays occur when a fraudster captures a high-resolution image of an ID from a mobile or video screen, rather than photographing the original document. Other fraudsters may photocopy a document and cut it to resemble a genuine ID, or cover parts of a document (like the portrait) with alternative, tampered data. Each of these techniques leaves a signature that liveness detection looks for. For example, moiré patterns will distort an on-screen image, consumer-grade printers produce a narrower range of colors than more advanced ID printing techniques, and tampered images may contain unusual lighting patterns or unexpected edges within a document.
Traditional document authentication vs. Document liveness detection
Where traditional document authentication focuses on the presence of document security elements like holograms or the consistency of the ID’s layout, document liveness verification uses more sophisticated image analysis and AI to look for evidence of genuine presence, and machine learning to identify elements that are commonly reused by fraudsters. Mitek’s IDLive Doc, for example, combines the use of several specialized AI models to detect the aforementioned ID artifacts and returns a confidence score in less than a second.
The benefits of document liveness detection
Document liveness detection has numerous fraud-fighting benefits and can be an important first line of defense. Based on recent in-house analysis by Mitek Systems, 90% of fraud in digital onboarding and KYC involves document fraud or presentation attacks, making the detection and prevention of these attacks a high-impact way to reduce overall losses.
There are several ways liveness detection improves security and operational processes for identity verification:
- Stops fraud at the top of the funnel: By filtering out spoofed documents before they progress through the identity verification flow, the risk of the account passing subsequent checks is removed.
- Reduces manual reviews: Flagging and rejecting questionable images early cuts down on the number of cases subsequently escalated for human analysis.
- Improves user experience: Document liveness checks run in the background, with no need for extra steps from the user.
- Enhances regulatory compliance: Strengthens digital KYC programs while also helping organizations meet AML and anti-fraud standards.
- Increases confidence in downstream checks: When an ID is verified as present and not tampered with, the reliability of subsequent face matching and data extraction can be trusted.
How document liveness detection works in identity verification
Document liveness detection is typically integrated directly into the remote identity verification flow using a four-step process:
Capture
Using their mobile phone or webcam, the user captures an image of both sides of their government-issued identification. The system generally provides guidance to help users frame both the front and back of the image correctly and to avoid glare that might result in an image being rejected. The image captured here is used both for liveness analysis and for document verification.
Assessment
Behind the scenes, AI, machine learning models, and deep neural networks analyze the captured image across numerous data points. These algorithms attempt to determine if the genuine ID document is physically present and unaltered, often detecting common fraud indicators such as moiré patterns, unusual lighting, or inconsistent edges.
There are often multiple models working simultaneously to evaluate the image and return a unified liveness score. Mitek has implemented its proprietary Document Liveness algorithms to deepen its analysis capabilities and provide further image analysis, including image angles, lighting, pixelation, and thousands of other image variations that may indicate a document is not physically present or has been altered.
Data extraction
Once the document has passed the liveness check, identity data is extracted from it by the use of optical character recognition (OCR). Name, date of birth, the ID number, and other information are extracted from the printed fields and barcodes on the document. While data extraction is not part of the liveness detection process, it relies on liveness detection to ensure the image being processed at this stage is genuine and unaltered.
Document verification
Once these steps are passed, the extracted data and document image are all evaluated for authenticity. In this step, the check includes elements like types of fonts, layout, barcode consistency, and whether extracted data is consistent with expected templates. For documents such as passports that support it, this may also include NFC chip reading and comparison. Used together, liveness and document verification provide layers of protection. Liveness confirms that the document is present and not tampered with. Document verification confirms that it is valid.
How document liveness detection prevents identity theft and fraud
Document liveness detection helps prevent identity theft and fraud by detecting spoofed or non-genuine ID captures, such as printed copies, screen replays, or physically altered documents. Liveness detection is especially effective when it comes to catching the most common types of document fraud, like:
- Screen replay attacks, where an image of an ID is displayed on a phone or tablet
- Printed copy attacks, which use printouts of real IDs on paper
- Portrait substitution attacks, where an ID has been physically (not digitally) altered to cover or replace the photo of the ID holder with the fraudster’s photo
By detecting these spoofing attacks, which make up the majority of the ID fraud in the digital onboarding process, liveness detection is able to block fraudulent applications before they can proceed to steps like biometric matching or even account approval, stopping fraudulent account creation at its inception.
Long-term, this protects institutions against the financial losses and reputational damage associated with fraud. It also prevents bad actors from gaining access to funds, services or platforms under false identities, whatever their specific intent.
Use cases for document liveness detection in ID verification
Document liveness is an important layer of protection in any remote identity proofing process. Some common use cases are:
- Banking/ Fintech: Onboarding users on any platform with KYC requirements, especially when issuing credit or providing access to financial services
- Cryptocurrency exchanges: Verifying customers during account setup or before high-value transaction
- Online marketplaces or gig work: Verifying sellers on peer-to-peer ecommerce platforms, renters and/or landlords, service providers (Uber/ Doordash)
- Gaming or gambling: Confirming age and/or identity, in compliance with local laws
- Public sector: For digitally verifying identity when applying for government benefits, health services, or licenses, or for controlling border access
- Education: Confirming the identity of a person taking an online course or exam or for student loan applications
- Remote hiring and contractors: For verifying employment eligibility or ensuring employees are in a specific region
Limitations of document liveness detection
Liveness detection is a powerful layer of protection, but it can’t solve every identity challenge alone. For example, liveness detection cannot verify that the person submitting the document is actually the rightful owner of the document. There are also edge cases that limit the effectiveness of liveness detection algorithms. For example, low-quality images may lack enough data to identify fraud. And some older documents may also be printed onto paper or contain portraits that are physically attached, both of which may trigger fraud alerts.
However, these issues can be mitigated. Institutions that work with documents that might regularly trigger these checks should look for a vendor like Mitek that allows tuning of detection thresholds or disabling of specific checks across regions or document types.
How to know if document liveness is right for your business
Document liveness detection is especially useful not just for businesses in the industries referenced above, but any business that:
- Verifies user IDs remotely
- Operates in a high-risk, highly regulated industry
- Needs to reduce false positives and fraud-related costs
- Has a heavily manual account approval process but is scaling quickly and needs to automate manual ID reviews
For organizations that are already collecting ID images during onboarding or verification, liveness detection is generally simple to add as a background process. If you’ve seen an increase in fraud attempts utilizing fake and/or reused ID images, the implementation of liveness checks can dramatically reduce the efficacy of those attempts.
How to implement document liveness detection
Document liveness solutions are designed for easy integration into your existing verification workflow. Providers like Mitek offer flexible deployment options, such as cloud-based APIs, SDKs, or containerized options that can be deployed in private environments.
The process works in the background of your existing verification, using the same document image you’ve already captured for OCR, and delivers results in real-time or near real-time. Importantly, Mitek utilizes passive liveness detection, meaning a fraudster won’t even be able to tell if a liveness check is going on or what is being analyzed. This makes it more difficult for them to attempt to reverse-engineer the process and avoid detection.
Steps in a typical implementation include integrating the liveness detection SDK or API into your existing platform, configuring your risk threshold and fallbacks, and a testing phase with samples of valid and altered printed/screen-based images. Then, the implementation is monitored and fine-tuned over time.
Depending on your infrastructure, implementation can take as little as a few days to complete. However, some organizations choose to start with a proof of concept first to validate impact.
What to expect in the future
Document liveness analysis is a rapidly evolving area, with regular advancements in AI software, hardware, and training improving the ability of deep neural networks to detect anomalies and threat patterns. Across the industry, future developments are expected to include:
- Greater use of multimodal analysis, combining document liveness analysis with biometric and behavioral signals
- Expanded capabilities for analysis of digital and chip-based IDs
- Stronger industry standards and third-party analysis benchmarking
- Continual evolution to provide protection against advanced spoofing methods like synthetic IDs and deepfake-style documents
Liveness detection is a first line of defense in identity verification, and an important anti-fraud tool that allows organizations to continue to deliver seamless digital customer onboarding experiences while improving compliance and mitigating fraud.
See how Mitek’s document liveness detection works in action.