We Built ImageDeep Because
False Alarms Kill Security.
Traditional CCTV generates thousands of alerts — almost all false. Guards stop paying attention. Real threats get missed. We set out to fix that by combining edge AI with human verification as an optional upgrade, so the alerts you receive are real threats, verified in seconds.
Irish Roots. Global Reach.
ImageDeep was founded in Quarry Bay, Hong Kong, by an Irish engineering team — giving us roots in two of the world's most trusted data jurisdictions. Hong Kong is our Asia-Pacific operations base; Ireland is where our compliance and data-protection DNA comes from. Together they let us serve enterprise clients across Asia, Europe and the Americas to one consistent standard.
From that Hong Kong base we monitor sites across the region, while our Dublin office anchors GDPR-native operations in Europe. European and US clients benefit from NDAA-compliant hardware and region-specific data residency. We understand the regulatory landscape, the operational challenges, and the cultural expectations of security in every market we serve.
Today, we protect sites across construction, retail, healthcare, education, and energy infrastructure — from Hong Kong and Singapore to the UK, Ireland and the United States — with the same promise we made on day one: human-verified alerts as an optional upgrade, delivered in seconds.
Before AI was fundable,
we were already doing this.
ImageDeep started before "physical AI" was a term anyone used. The founding team's background is in machine vision going back over 25 years — PCB inspection systems for pharmaceutical and medical device manufacturers, smart city deployments, airport operations. That's where the human-verified alert standard comes from: in a factory line or an airport, a false alarm isn't an inconvenience, it stops production or grounds people. We built ImageDeep on the same standard, applied to live video security.
Beyond analysing video, ImageDeep continuously verifies the health of the entire underlying infrastructure — every camera, power source, network link, and recording device — so the system watching your site is itself always being watched. Most platforms assume the cameras are working. We prove it, in real time, with the same human-verified alert process — a key differentiator that eliminates blind spots, lost evidence, and finger-pointing.
That's still what separates us from most of what's being built under the "AI security" label today. A lot of it is new — untested outside a demo, running on general-purpose models never designed for this. Ours has been running on live construction sites across Hong Kong's biggest developments for years, with trained human operators checking actionable events under optional 24/7 operator monitoring before a single alert reaches a client. The AI doesn't replace judgment. It gives the right person the right information fast enough to act on it.
Three patents sit behind the detection work. The company has been through Hong Kong's I&T funding programme. Clients include Gammon Construction, PERI Group and Harrow International School — real deployments, not pilots.
What's changed is the industry catching up to the idea that AI in the physical world has to be reliable, not just impressive. We'd argue we never needed the wave to catch up — but we're glad the rest of the market is finally asking the right question: not "can it detect something," but "can you trust it enough to act on it."
Built in the field.
Not in a boardroom.
Solar-powered, off-grid, watching a remote perimeter within hours of arriving on site.
We work closely with camera manufacturers, security & safety integrators and distributors
ImageDeep goes to market hand-in-hand with local integrators and distributors. You own the client relationship and the installation — we provide the AI + human-verification intelligence layer, the 24/7 monitoring centre, and the co-marketing support to help you win and retain more business. It's a core part of how we grow across Greater China, Southeast Asia, Europe and the Americas.
Explore the Partner Program
We own the whole stack
From PCB design to the software in our monitoring centre, we build it ourselves — and we design custom hardware for clients when off-the-shelf won't do. Components and manufacturing are sourced through our sister company, ImageDeep Sourcing.
From reducing false alerts
to removing the noise.
Niall Dorr spent years building the computer vision systems behind a leading global proactive video monitoring provider, engineering the false-positive reduction technology that made remote monitoring viable at scale. He founded ImageDeep to go further — from reducing false alerts toward removing them.
Connect with Niall on LinkedInNeed a custom enterprise solution? (Hardware/Software)
Speak to our engineering team directly on WhatsApp.
Message our engineers on WhatsApp
Use your phone and chat now
Point your WhatsApp camera at the code to get our number and start the chat.
Our Journey
ImageDeep founded in Hong Kong to solve the false alarm epidemic in video surveillance.
Hybrid AI + human verification model deployed across first construction and retail sites in Hong Kong and the Asia-Pacific region.
Platform expanded to thousands of camera models, 17 AI agents, and multi-site enterprise dashboards.
Trusted by Gammon, PERI and Harrow — protecting sites across Hong Kong and Europe.
The Hybrid AI + Human Model
AI never blinks. Humans never guess. Together, they deliver security that actually works.
Human-Verified Alerts
With optional 24/7 operator monitoring, a trained operator checks actionable events before they reach you. No noise, fewer false alarms — only real threats, verified in seconds.
Live in 48 Hours
We deploy on your existing IP cameras. No rip-and-replace, no construction work. Most sites are fully operational within two business days.
Human Verification, in Writing
With 24/7 operator monitoring, how actionable events are verified and escalated is written into your contract.
What We Stand For
Our values shape every product decision, every deployment, and every alert we send.
Precision Over Noise
We built our entire model around fewer false alerts. With optional 24/7 operator monitoring, a trained operator checks actionable events before they reach you.
People First
Security is about protecting people. Our human verification layer ensures every response is measured, accurate, and responsible.
Global Reach, Local Roots
Founded in Hong Kong, with a European office in Dublin, serving clients across North America, the UK & Europe, the Middle East and Asia-Pacific. One platform, one standard, everywhere.
The Expertise Behind ImageDeep
Our team brings decades of combined experience in security operations, AI engineering, and enterprise compliance across Asia Pacific.
Security Operations
Former military, police, and enterprise security leaders with 15+ years in physical security across Asia Pacific.
AI Engineering
PhD-level computer vision and ML engineers building edge AI that runs on thousands of camera models with sub-2-second detection.
Compliance & Governance
HK Security Bureau licensing, GDPR, and NDAA specialists ensuring every deployment meets regulatory and audit requirements.
Client Success
Follow-the-sun support across Hong Kong, Ireland and the US — native-language operators covering APAC, EMEA, the Middle East and the Americas, 24/7.
Real Results, Real Clients
We measure success in incidents prevented, costs saved, and hours returned to our clients.
Certifications & Compliance
Enterprise-grade security requires enterprise-grade trust. Our certifications speak for themselves.
In the press
Ready to See It in Action?
Book an assessment. We'll review your site, check your camera compatibility, and show you exactly how ImageDeep would work for you.