X-Ray FAQ : Accuracy, Comparison to Human Detection, and Model Training Process

How accurate is the AI at detecting objects?

The AI is specifically trained to identify high-risk items based on established threat models. While highly advanced, no automated system is 100% foolproof. Detection accuracy can be influenced by:

• Obstruction: Items densely packed or layered behind heavy shielding (like lead or thick steel).

• Distortion: Objects positioned at extreme angles that obscure their identifiable shape.

• Environmental Factors: The complexity of the bag's contents.

The system is designed to act as a powerful "second set of eyes," flagging anomalies that require closer inspection.

How does AI compare to manual human detection?

AI significantly enhances security by providing speed and consistency that human operators cannot maintain over long shifts.

• Fatigue-Free: Unlike humans, AI does not suffer from "search fatigue" or distraction; its detection threshold remains identical from the first bag of the day to the last.

• Processing Speed: AI analyzes complex images in milliseconds, allowing for a faster throughput of bags and shorter wait times.

• Consistency: The AI applies the same rigorous logic to every scan, eliminating the subjective nature of manual human review.

Can the AI be trained to recognize new types of objects?

Yes. Our AI model is dynamic and constantly evolving.

• Current Process: If you have specific items unique to your environment that need to be detected, you can scan those items and provide the images to our engineering team. We use this data to update and refine your local model.

• The Future of Training: In an upcoming app release, we are streamlining this process. Operators will be able to tag an object directly on the iPad interface with a single button, automatically sending the image to the model for future training and optimization.

Quick Comparison: AI vs. Human Inspection