Solution Overview
DEXMA — AI Vision Recognition
Deep learning-based visual intelligence system
DEXMA is Gmission's cutting-edge AI Vision Recognition System (AI VRS) that delivers accurate, real-time identification using advanced deep learning models. From facial recognition for access control to license plate recognition for traffic management, DEXMA provides enterprise-grade visual intelligence solutions that integrate seamlessly with existing security infrastructure and IoT ecosystems.
Key Features
Face Recognition
Sub-second facial identification against watchlists of millions of enrolled identities, with liveness detection to defeat spoofing attacks using photos or video.
License Plate Recognition
Reads vehicle plates across 50+ country formats in motion at highway speeds, powering toll collection, parking automation, and law enforcement workflows.
DEXMA AQUA
Vision-based water quality monitoring that detects turbidity, algae blooms, and pollution events through camera feeds without requiring chemical sensor replacement cycles.
Object Detection & Tracking
Multi-class object detector identifies people, vehicles, packages, and custom-trained assets, maintaining persistent identity tracks across camera handoffs in large-area surveillance.
Anomaly Detection
Unsupervised baseline learning flags deviations from normal scene behavior — abandoned objects, loitering, crowd density spikes — triggering alerts before incidents escalate.
Edge & Cloud Hybrid
Inference runs on-device for latency-critical decisions, with encrypted metadata synchronized to the cloud for centralized management, audit logging, and model updates.
Open Integration API
REST and WebSocket APIs connect DEXMA results to VMS platforms, access control systems, ERP, and custom applications with pre-built connectors for leading vendors.
Privacy-by-Design
Configurable PII masking, on-device data minimization, and role-based access controls satisfy GDPR and Korean Personal Information Protection Act requirements out of the box.
Key Benefits
Recognition Accuracy
Top-1 identification accuracy on controlled access-control benchmarks, maintained under variable lighting and partial occlusion conditions.
Inference Latency
End-to-end time from camera frame capture to alert delivery on edge hardware, enabling real-time response without cloud round-trips.
Security Incident Reduction
Reduction in unauthorized access events reported by enterprise customers within the first six months of DEXMA deployment.
Operational Cost Savings
Reduction in manual security guard hours through AI-assisted monitoring that prioritizes genuine alerts over false positives.