FaceOnLive — Face Recognition SDK for Linux

Server-side face detection and 1:1 / 1:N face matching for Linux, exposed through a Python interface over the native engine. Runs entirely on your own server.

Part of the FaceOnLive on-premises biometric SDK suite.

📦 Download the full SDK: this repository contains the source/demo code only. The complete SDK — engine libraries and model files, with the full project structure — is attached to the Releases page. Download the archive and extract it over this project.

Features

• Face detection and biometric template extraction.
• Face similarity comparison for verification and identification.
• On-premises and offline; suitable for backend / batch pipelines.

Requirements

OS	Linux x86-64
Runtime	Python 3.8+
Engine	libttvfaceengine (included under engine/)

Setup

1. Get a license key — free trial at https://faceonlive.com.
2. Provide the key via the LICENSE_KEY environment variable, or place it in license.txt (replace the <YOUR_LICENSE_KEY> placeholder).
3. Install Python dependencies and run the provided demo.

Quick start (Python)

from engine.header import *   # exposes the bindings below

init_sdk()                                  # ttv_init
t1 = extract_template(image1)               # ttv_extract_feature
t2 = extract_template(image2)
score = calculate_similarity(t1, t2)        # ttv_compare_feature  (0.0 – 1.0)

API reference (Python bindings → native)

Binding	Native	Description
get_version()	ttv_version	SDK version.
get_deviceid()	ttv_get_hwid	Machine hardware ID (for offline licensing).
init_sdk() / init_sdk_offline()	ttv_init / ttv_init_offline	Initialize online or offline.
extract_template(image)	ttv_extract_feature	Extract a face template.
calculate_similarity(t1, t2)	ttv_compare_feature	Compare two templates (0.0–1.0).

License & support

Requires a valid license key — get one at faceonlive.com. Keep license.txt out of version control. Questions: contact@faceonlive.com