Fingerprint capture and matching with no scanner. Finger AI — deployed for law-enforcement as "ThumbsUp" — turns an ordinary phone camera into a contactless fingerprint station: snap or sweep a fingertip, the Cams extractor builds a standard ISO/IEC 19794-2 template, and the Fingerprint Match Engine enrolls it, searches it 1:N, and charts the matching minutiae side by side for a human to verify.
Overview
Finger AI is a Flutter Android app for contactless fingerprint (fingerphoto) capture, enrollment and identification. The camera takes a fingertip image — a quick photo or a multi-frame panoramic sweep the server mosaics into a maximum-ridge-area print — behind an on-device quality gate (sharpness, brightness, glare), tap-to-focus, flash/zoom and voice commands. The Cams fingerprint extractor converts the capture into an ISO/IEC 19794-2 minutiae template at ~500 dpi and returns an enhanced grey with the minutiae and singular points drawn on it; the Fingerprint Match Engine then runs 1:N identify, 1:1 verify, and a detailed 1:1 compare that numbers the matching minutiae on both prints. Operators can enroll a person's ten fingers, review / crop / retake / remove each print, load an existing person by ID to edit them, and page through the closest candidates even when a search falls below threshold. One app serves two audiences: hygienic, hardware-free attendance on employees' own phones, and forensic contactless fingerprint identification for police departments — with results delivered over the standard Cams REST + callback pipeline to your server, ERP, or a review portal.
Organisations wanting hygienic, hardware-free fingerprint attendance on employees' own phones — and police / forensic teams needing contactless fingerprint capture with ISO-standard minutiae, 1:N search and examiner-grade side-by-side comparison.
Capabilities
Snap a single fingertip photo, or sweep the phone around it for a multi-frame panorama the server mosaics into a maximum-area print — no scanner, no contact.
Sharpness, brightness and glare are checked on the phone before upload, so a blurry or washed-out frame is caught with a retry hint instead of a bad match.
The fingerphoto becomes a standard ISO/IEC 19794-2 minutiae template at ~500 dpi, with border artifacts suppressed so only genuine ridge features remain.
Templates are searched against your gallery and scored by the Cams Fingerprint Match Engine, returning the best match above a configurable threshold plus the forensic point count.
A detailed 1:1 compare shows the probe and a candidate side by side with numbered matching minutiae — and lets an officer page through the closest candidates even below threshold.
Enroll a person's ten fingers, review / crop / retake / remove each print, or load an existing person by ID to edit their prints and save the changes.
Contact-free capture on people's own phones — hardware-free fingerprint attendance for shared and clinical environments.
Standard Cams REST with serial + auth-token and optional AES-256 encryption; results delivered over the Cams callback pipeline to your server, ERP, or review portal.
Features
Grab a key, read the docs, and ship. Our team helps with your first integration.
FAQ
No. It captures a contactless fingerphoto with the phone camera — a single photo or a panoramic sweep — and the server converts it to an ISO 19794-2 template. A clip-on macro lens sharpens ridge detail but no dedicated sensor is required.
Yes. Every capture is extracted into an ISO/IEC 19794-2 minutiae template at ~500 dpi — the same format the whole Cams fingerprint engine uses — so it interoperates with existing enrollments.
Yes. Beyond the automatic 1:N result, the app opens a side-by-side view of the probe and a candidate with the matched minutiae numbered on both prints and the point count shown, and it can page through the closest candidates even when the search is below threshold — so a human makes the final call.
A photo is one quick frame; a panoramic sweep takes several frames as you move the phone around the fingertip and the server mosaics them into a single larger-area print with more ridge detail and minutiae — useful for enrollment and difficult fingers.
Contactless fingerphoto matching is inherently lower-overlap than a pressed contact print, so scores run lower and the threshold is calibrated for fingerphoto rather than borrowed from contact sensors. Panoramic capture, the on-device quality gate, ~500 dpi normalisation and border-artifact suppression exist to make genuine matches reliable and human-verifiable.
Over the standard Cams REST API and RealTime callback pipeline — to your own server or ERP, or to a review portal such as the police crime-review portal — with retry/backoff so nothing is lost if the destination is briefly down.
Yes. Load a person by their user ID to preload their fingers, then retake or remove individual prints, add new ones, and save — removals are applied on save.
Related
A tablet-PC face-attendance kiosk with on-device recognition, liveness detection, an offline queue, and real-time MQTT push.
ExploreAPPPhone-based IN/OUT attendance for field and remote teams, with fast two-QR device provisioning.
ExploreAPIConnect any internet-enabled biometric attendance or access-control device directly to your web application over HTTP — no local software, PC, or database required.
ExploreWe use cookies to measure traffic and improve BiometricAPI.com. Analytics stay off until you accept. Learn more.