Passive liveness.
As simple as a selfie.
IDmission's AI-powered passive liveness detection. No blinking. No smiling. No head turns. Just a single selfie — and a 160-layer deep neural network that instantly determines if the person is real, physically present, and who they claim to be.
Three layers of proof in a single selfie
Each selfie is evaluated across three dimensions to establish identity with confidence.
Person
Is this a real human being — not a statue, mask, face bust, or virtual persona posing as a person?
Physically Present
Is this person authenticating in real time — not using a photo, video, deepfake, or replay attack?
Precisely Who They Claim
Is this person who they say they are — and do they have the right to access the service requested?
Passive Liveness
Remove hat & glasses
Completely passive verification
No commands. No friction. The user simply takes a selfie and the CNN does the rest — in under a second.
Remove hat & glasses
The SDK follows ICAO guidelines and prompts the user to remove obstructing accessories for optimal capture quality.
Fit face in the oval
Real-time guidance checks face fit (65–100% tolerance), head angle (±10° Y, ±5° Z), eye openness, lighting, and saturation.
Auto-capture & AI analysis
Once all criteria pass, the selfie is captured automatically. The 160-layer CNN analyzes the image and returns a real/spoof determination with a confidence score — requiring a real score above 0.9.
Result in milliseconds
50 ms on-device latency. 600 ms end-to-end response time. The user never waits — the entire process feels like taking a normal selfie.
Explore the Data
Go beyond the surface—drill down into the specific model precision metrics and our document realness performance benchmarks.
Trained against 15 attack types
The CNN is trained with over 500,000 balanced real and spoof images, then validated against over 50,000 test images — covering every known presentation attack vector.
Attack Types Defeated
Available everywhere you build
IDmission's AI-powered passive liveness detection works across every integration channel — native mobile, web, or server-side API.
Android SDK
Native Kotlin/Java SDK with on-device TFLite models.
iOS SDK
Native Swift/ObjC SDK with CoreML inference.
WebSDK
JavaScript library with in-browser ML models. Any browser.
REST API
Server-side integration for custom capture pipelines.
Stop spoofing. Start with a selfie.
Certified to ISO 30107-3 Level 2. 99% accuracy. 50 ms latency. Zero user friction. See IDmission's AI-powered passive liveness detection in action.