Computer Vision
AI that sees, understands, and acts.
From perception to reasoning. Our computer vision systems combine recognition, contextual analysis, and semantic understanding to enable real-world intelligence.
AI Scan Results
Live recognition feed
Perception Layer
Visual Detection & Recognition
Our vision systems identify and classify objects, brands, and vehicles in real time. From fleet management to intelligent access control, these models form the foundation of visual automation.
- ALPR Fleet System
A vehicle and brand recognition platform used in a live workshop in Ireland. It detects license plates, car makes, and models to enhance fleet management and workflow efficiency.
Context Layer
Scene Understanding & Monitoring
We build systems that interpret the dynamics of real-world environments — analyzing movement, density, and contextual states. From camera streams to actionable intelligence.
- Traffic control room
Developed using NVIDIA TrafficCamNet and DeepStream. This live demo monitors traffic load and barrier status at a ferry terminal in Norway, delivering real-time analytics and visual overlays.
Live reasoning session
Q: Why did the forklift stop?
A: AF1 lost a cardbox box while turning left.
Reasoning Layer
Vision-Language Integration
Exploring the connection between visual perception and language understanding. These experiments bridge visual models with reasoning capabilities — enabling AI to see, describe, and act.
- Reasoning Warehouse System
A reasoning model observes a robotic warehouse and explains its decisions. Cosmos Reasoning 2 analyzes warehouse footage to explain a stop event and infer next actions.
Ghostblitz AI
An AI agent that plays Ghostblitz by recognizing objects in real time.
NVIDIA Inception Program
Noctis AI has been accepted into the NVIDIA Inception program, an initiative designed to support startups that are revolutionizing industries with advancements in AI and deep learning.
As a member, we gain access to NVIDIA's ecosystem of technical resources, including GPU-accelerated infrastructure, research toolkits, and education platforms such as NVIDIA Cosmos, to advance our exploration of world models, computer vision, and intelligent automation. This connection empowers us to push the boundaries of perception, reasoning, and real-time AI systems.