Scaling Face Search (1:N) with the Didit Android SDK

Seamless IntegrationIntegrate powerful 1:N Face Search directly into your Android applications using Didit's native Kotlin SDK, simplifying complex biometric matching processes.

Robust Fraud PreventionProactively identify duplicate accounts and prevent fraud by comparing new user biometrics against your entire database of verified identities.

Configurable SecurityCustomize facial similarity thresholds and manage automatic decline conditions like NO_FACE_DETECTED or FACE_IN_BLOCKLIST to fit your specific risk profile.

Developer-First ApproachDidit offers a modular, AI-native platform with clean APIs and a free core KYC tier, empowering developers to build scalable and secure identity solutions efficiently.

The Power of 1:N Face Search in Mobile Applications

In today's digital economy, verifying user identities and preventing fraud are paramount for businesses across all sectors. As mobile applications become the primary interface for user interaction, the need for robust, on-device identity verification solutions has never been greater. One of the most effective tools in this arsenal is 1:N Face Search, a biometric capability that allows you to compare a new user's face against a database of existing verified identities. This process, often referred to as 'one-to-many' matching, is critical for detecting duplicate accounts, combating synthetic identity fraud, and ensuring the integrity of your user base.

Integrating such advanced functionality can be complex, requiring deep expertise in computer vision, biometric analysis, and secure data handling. However, with the right tools, developers can seamlessly embed these capabilities directly into their Android applications, providing a frictionless user experience while maintaining high security standards. This is where Didit's Android SDK shines, offering a streamlined path to implementing scalable 1:N Face Search.

Understanding Didit's Face Search Capabilities

Didit's Face Search feature is designed to be a powerful fraud prevention tool. It enables organizations to search for a specific face across all previously approved identity verification sessions. This means that when a new user attempts to register or perform a sensitive action, their biometric data can be instantly cross-referenced with your entire database of verified users. The system generates similarity scores, indicating the likelihood of a match, and applies configurable thresholds to determine if a match should be returned.

Key aspects of Didit's Face Search include:

• Duplicate Account Detection: Identify users attempting to create multiple accounts, a common tactic for fraud and policy circumvention.
• Blocklist Integration: Automatically check new faces against a blocklist of known fraudulent individuals, preventing them from accessing your services. Didit's system includes automatic decline conditions like FACE_IN_BLOCKLIST.
• Configurable Sensitivity: Adjust the similarity threshold (typically around 70%) to balance between accuracy and the number of potential matches. Higher thresholds reduce false positives but might miss some matches, while lower thresholds increase the net.
• Handling Multiple Faces: Configure whether your application allows or disallows multiple faces in the search image. If disallowed, the session will fail, ensuring clear, unambiguous searches.

These capabilities are crucial for maintaining a secure and trustworthy platform, protecting both your business and legitimate users from malicious actors.

Integrating Face Search with the Didit Android SDK

Integrating Didit's Face Search into your Android application is a straightforward process thanks to the developer-first design of the Didit Android SDK. This native Kotlin SDK provides all the necessary components for a smooth integration, including camera handling, liveness detection, and NFC verification, alongside the core ID Verification and Face Match functionalities.

The integration typically involves:

1. SDK Installation: Adding the Didit SDK to your project's Gradle dependencies, including the necessary Maven repository. The SDK handles required permissions like INTERNET, CAMERA, and optionally NFC, merging them into your app's manifest.
2. Initiating a Session: Starting a Didit verification session where the user provides their face image. This image is then used for the 1:N Face Search.
3. Processing Results: Receiving the Face Search report, which is a JSON object containing details about any matches found. This report includes session_id, similarity_percentage, user_details (partially masked for security), and a temporary match_image_url.
4. Handling Warnings and Errors: The SDK provides detailed warnings (e.g., NO_FACE_DETECTED, MULTIPLE_FACES_DETECTED) to ensure high-quality input and reliable results.

The SDK abstracts away much of the complexity, allowing your development team to focus on building a great user experience rather than reinventing the wheel for biometric processing. The temporary URLs for match images expire after 60 minutes, reinforcing security best practices by encouraging applications to store only verification statuses and similarity scores, not raw biometric data.

Best Practices for Scalable Face Search Implementation

To maximize the effectiveness of 1:N Face Search and ensure scalability, consider these best practices:

• Optimize Image Capture: Guide users to provide clear, well-lit images with a single face. Didit's Liveness Detection, integrated within the SDK, helps ensure the image is of a real, present person and not a spoof attempt.
• Set Appropriate Thresholds: Continuously monitor and adjust your similarity thresholds based on your risk tolerance and the specific use case. What works for low-risk scenarios might not be sufficient for high-value transactions.
• Automate Review Workflows: Leverage Didit’s orchestrated workflows to automate responses to Face Search results. For instance, automatically decline sessions with a high similarity match to a blocklisted individual, or flag sessions with moderate matches for manual review.
• Privacy and Data Security: Always prioritize user privacy. Ensure that biometric data is handled in compliance with relevant regulations (e.g., GDPR, CCPA). Didit's architecture is designed with security in mind, providing temporary image URLs and encouraging minimal data retention.
• Monitor Performance: Regularly review Face Search reports and warning logs to identify patterns, improve processes, and fine-tune your configuration. Warnings like LOW_QUALITY_SEARCH_IMAGE can indicate areas where user guidance or image capture processes can be improved.

By following these guidelines, you can build a highly effective and scalable fraud prevention system that leverages the full power of biometric identity verification.

How Didit Helps

Didit provides an AI-native, developer-first identity platform that makes scaling 1:N Face Search straightforward and efficient. Our modular architecture allows you to plug-and-play identity checks, including advanced Face Search, 1:1 Face Match, and Passive & Active Liveness detection, directly into your applications. The Didit Android SDK offers a seamless integration experience for mobile developers, enabling them to embed robust identity verification without extensive effort.

Didit stands out by offering Free Core KYC, meaning you can start verifying identities without upfront costs. Our pay-per-successful-check model and absence of setup fees make it an economically viable solution for businesses of all sizes. The platform's AI-native capabilities ensure high accuracy and continuous improvement in biometric matching, while our clean APIs and comprehensive documentation empower developers. With Didit, you can orchestrate complex risk workflows, automate trust, and focus on your core business, knowing your identity verification is handled by a leading-edge, global solution.

Ready to Get Started?

Ready to see Didit in action? Get a free demo today.

Start verifying identities for free with Didit's free tier.