trm logo

Scylla Facial Recognition: Automated Identity Detection

An AI-powered system that builds a dynamic, real-time facial database to enable access control, visitor monitoring, and analytics — all without manual intervention.

Scylla's Auto-Enrollment

How it works:

● Scylla Facial Recognition with Automatic Enrollment is a fully AI-driven solution that manages the entire process. It eliminates the need for manually creating watchlists — though that option remains available.
● Unlike traditional systems that rely on static lists, Scylla AI automatically builds a dynamic facial database. Each face captured by the cameras is assigned a unique real-time identifier, allowing accurate tracking of repeat visitors.
● Scylla stands out from other providers by offering an automatic enrollment window of 1 to 1.5 months, while most competitors limit it to just a few days.

Scylla's Auto-Enrollment supports powerful business analytics features:

  • Peak hours and dwell time

  • Returning visitors

  • Emotional expression analysis

  • Number of new visitors

Scylla's Auto-Enrollment

Discover the power of facial recognition with Scylla’s groundbreaking Auto-Enrollment feature:

Real-Time Database Creation

Scylla autonomously builds a database of detected individuals with their identification data, eliminating the need for manual data entry and management.

Every person appearing on a surveillance camera is automatically stored with a unique, traceable identifier. This transforms both security monitoring and visitor analytics by generating a living, evolving database over time.

Scylla's Auto-Enrollment

Powerful Forensic Search

Scylla’s historical search feature allows users to upload an image of a person and conduct a comprehensive search across previously collected data. This powerful feature leverages Scylla’s Auto-Enrollment system, which continuously generates and updates a database of unique anonymous identifiers. You can retrieve historical interaction data such as location tracking, visit frequency and patterns, behavior analysis, and anomaly detection.

Scylla's Auto-Enrollment

Consistent Identification for Situational Awareness

When an unknown individual reappears in the surveillance system, Scylla’s AI instantly recognizes them using a previously assigned unique identifier, enabling continuous monitoring and tracking across multiple visits or appearances.

This feature enhances security and also expands opportunities for customer engagement and personalization.

Scylla's Auto-Enrollment

Deeper Understanding of User Needs

In customer-focused environments, Scylla provides a powerful advantage: the ability to understand individual behavior, patterns, and interactions — even without prior customer records.
This empowers retailers, event organizers, and service providers to personalize experiences and gain insight into the preferences of returning visitors over time, all without collecting personal information.

Scylla's Auto-Enrollment

Extended Timeframes

Scylla’s Auto-Enrollment remains active for an extended period — from 1 to 1.5 months. This prolonged window is especially beneficial in scenarios where individuals need to be tracked or identified over longer durations.

FAQ

  • The facial recognition system operates with two types of databases:

    Vector Database – This stores generated facial vectors, which are irreversible. In other words, the original image cannot be reconstructed from these vectors.

    Image Database – This stores the actual facial images.

    Storing images is optional. If desired, the facial recognition system can function solely using the vector database, in full compliance with privacy preferences and data protection policies.

  • Yes, a pilot implementation can be conducted.

    Success will be measured using statistical data collected through the BI dashboard and advanced search features, providing clear insights into system performance and effectiveness during the trial period.

  • For Full HD video streams, Auto-Enrollment requires 500 PPM (pixels per meter), which typically corresponds to a detection range of about 2 meters (with better performance at shorter distances).

    For subsequent recognition after enrollment, only 250 PPM is needed, allowing for a typical recognition range of about 4 meters.

    The detection latency is approximately 3 to 4 seconds, depending on specific system configurations.

  • Absolutely. Scylla does not store any data that can be considered personally identifiable. We do not retain video recordings or images.

    The only data stored are alarm notifications, and their retention period can be adjusted according to the policies defined by the client.

  • The number is configurable and depends on the system’s hardware.

    Increasing the number of faces analyzed simultaneously requires more powerful hardware to ensure optimal performance.

  • Yes, Scylla’s facial recognition system remains effective even with significant transformations or changes in appearance.

    However, for re-identification to work, the individual must have been successfully enrolled in the database. During the enrollment phase, stricter criteria are applied — partially obscured faces are skipped to ensure high-quality registration in the system.

  • Scylla’s watchlist-based facial recognition focuses on the precise identification of pre-registered individuals for immediate response. Enrollment is done manually and curated to recognize flagged persons in real time. This approach is ideal for airports, transportation hubs, law enforcement, and event security where identifying known individuals is critical.

    In contrast, Scylla’s Auto-Enrollment solution is designed for high-traffic environments like casinos, retail stores, and corporate campuses. It enables dynamic tracking and generates behavioral analytics data. The facial database is built automatically, without the need for manual data entry, making it highly scalable and efficient for ongoing visitor analysis and engagement.

Fill out the form to receive a personalized consultation for your IT infrastructure:

Дякуємо, ми отримали ваше повідомлення і звʼяжемось в найближчий час! :)


Can't send form.

Please try again later.