Secure3D-CV: A scientific framework for integrity assessment of 3D data with an external Python library

Three-dimensional (3D) perception is widely used in robotics, autonomous vehicles and medicine. These systems often operate autonomously and are vulnerable to data tampering or spoofing attacks. Existing work on 2D adversarial detection does not generalise well to the high-dimensional and structured nature of 3D point clouds.

We propose Secure3D-CV, a method that builds on OpenCV by adding a library for outlier detection and threat assessment in 3D scenes. The method transforms each scene ( S) into a feature vector ( x = φ(S)) via geometric, spectral and volumetric descriptors. A statistical anomaly score ( A ( S ) = | x − { x ¯ } | 2 ) measures the deviation of ( x) from a benign distribution ({ x ¯ }). A supervised or one-class classifier ( f _θ ( x)) estimates ( P( y = 1| x)) where ( y) denotes a safe scene. The final decision score ( R( S)) strategically combines ( A( S)) and ( f _θ ( x)) to classify the scene as safe or not safe.

On two datasets, simulated point clouds and depth maps containing 2,000 samples, and 100 anonymised CT volumes, we show that Secure3D-CV detects injected outliers, occlusions and adversarial perturbations with high true-positive rates (99.3% on synthetic data and 98.7% on CT data) and low false-positive rates (below 1%). We demonstrate that the hybrid ( R( S)) approach provides a significant advantage over purely statistical or machine learning-based methods.

Secure3D-CV is a lightweight extension to OpenCV that monitors incoming 3D data streams, extracts robust features and computes risk scores to flag unsafe scenes. The library provides Python bindings and can be integrated into robotics or medical-imaging pipelines. This work contributes a principled threat model for 3D data and demonstrates that combining anomaly scoring with a learned classifier fundamentally improves detection performance compared with baseline thresholding and state-of-the-art methods.

Imagine a self-driving car or a medical robot that uses 3D cameras (like LiDAR or depth sensors) to see the world. If a hacker or a system glitch subtly changes the data coming from these sensors—for example, by adding a fake object or slightly shifting a measurement—the car could crash, or the robot could make a critical mistake. Current computer vision systems (like those based on OpenCV) are great at processing 3D data, but they are often blind to these subtle, malicious changes. They focus on functionality, not security.

Secure3D-CV is a new, lightweight security layer designed to act as a digital bodyguard for 3D data streams. It works alongside existing computer vision pipelines (like OpenCV) without slowing them down significantly. Its main job is to constantly check incoming 3D data (like point clouds or depth maps) for signs of tampering, anomalies, or cyberattacks.

The core innovation of Secure3D-CV is its hybrid risk score. Instead of relying on just one method, it combines two powerful techniques:

• 1.

  Statistical Anomaly Detection: This is like a general alarm. It checks if the new data looks statistically "normal" compared to what the system usually sees. This is great for catching completely new, unknown threats or sensor failures.

• 2.

  Machine Learning Classifier: This is like a trained detective. It uses a model trained on known attack patterns to precisely identify specific, sophisticated cyberattacks.

By combining these two methods into a single, adjustable risk score, Secure3D-CV achieves much higher accuracy and robustness than any single method alone. Our tests show it can detect over 99% of attacks while keeping false alarms very low.

• High Reliability: The hybrid approach is proven to be superior to existing state-of-the-art methods, making autonomous systems safer.

• Low Overhead: It adds minimal processing time (less than 50% overhead), meaning it can be used in real-time applications like robotics and autonomous vehicles without performance issues.

• Easy Integration: It is delivered as a simple, open-source library that can be easily plugged into any existing OpenCV-based pipeline.

Secure3D-CV fills a critical security gap in 3D computer vision. It provides a practical, high-performance solution to ensure the integrity of 3D data, making the next generation of autonomous and medical systems more trustworthy and resilient against cyber threats.