Deep Learning Facial Recognition Engine

Enterprise Face Scan & Biometric Verification API

High-precision 512-dimensional facial embeddings, 1-to-1 face verification, and OpenCV Haar Cascade landmark detection with sub-200ms inference.

Explore API Documentation Live Health Status Download PDF Guide

Neural Backbone

Facenet

512-D L2-normalized vectors

Detector Engine

OpenCV / SSD

Haar Cascade fallback enabled

Distance Metric

Cosine

0.40 recommended threshold

Service Health

Healthy

Running on Ubuntu 26.04 KVM

Core API Endpoints

POST /represent

Extract facial embedding descriptor vector and facial coordinates from an image (base64 string or file path).

curl -X POST https://face.canscorp.com/represent \ -H "Content-Type: application/json" \ -d '{"img_base64": "data:image/jpeg;base64,...", "model": "Facenet"}'

Payload: JSON Response: 512-dim Float Array + Bounding Box Auth: Standard / Internal

POST /verify

Compare two facial images to verify whether they represent the same individual based on cosine distance.

curl -X POST https://face.canscorp.com/verify \ -H "Content-Type: application/json" \ -d '{"img1_base64": "...", "img2_base64": "...", "model": "Facenet"}'

Payload: JSON (img1 + img2) Response: verified (bool), distance (float)

GET /health

Inspect real-time service health, uptime, active model weights, and system metrics.

curl https://face.canscorp.com/health

Payload: None Response: {"status": "healthy", "service": "face-scan-api", "uptime_seconds": ...}