view_in_ar
VISITRAPLATFORM
Explore Demo
SYSTEM STATUS: READY
apiBACKEND: MOCK_SERVICE (FastAPI Contract v1.0)
speedENGINE FPS: 59.8
timerINFERENCE LATENCY: 16.4ms
NODE: GATEWAY-01
Inference Lab & Interactive Terminalv2.4.2

VISITRA RECOGNITION ENGINE

From multi-vector biometric enrollment to sub-20ms real-time identity matching. Decoupled client architecture fully configured for turnkey FastAPI deployment.

REGISTERED VECTORS1,842
CONFIDENCE THRESHOLD0.85 cos-θ
STEP 01

Register New Identity (Enrollment Interface)

Extract 512-D spatial embedding vectors across multi-angle photogrammetry

CONTRACT: POST /api/persons
VECTOR EXTRACTION TARGETARC-FACE v4 / ResNet100
Embedding Dim: 512-float323 of 4 Angle Vectors Extracted
Minimum 3 Slots Required
Front portrait
FRONT
512-D OK
0° Primary Portrait
Left Profile
LEFT-30°
512-D OK
Left Profile Key
Right Profile
RIGHT-30°
512-D OK
Right Profile Key
add_a_photoOPTIONALGlasses / Shadow
Variable Lighting
STEP 02

Stored Identity Registry

In-memory FAISS similarity vector index synced with mock repository

TOTAL IDENTITIES: 0 ACTIVEGET /api/persons
STEP 03

Real-Time Recognition Console

Interactive inference viewport connected to active pipeline state machine

DETECTING LANDMARKS
LIVE // CAM_01_GATE_NORTH
Live Match Telemetry Readout
STATUS: STANDBY INFERENCE

Awaiting subject approach. Use the simulation controls above to test real-time recognition state transitions.

PIPELINE: ACTIVEMODEL: DeepFace ResNet100 (Mocked)
Recent Recognition Telemetry Log
TIMESTAMPSUBJECT / NAMEIDENTIFIERCOSINE SCORESTATUSTERMINALLATENCY
Telemetry log cleared. Awaiting next recognition inference event.