Trueface Visionbox

State of the art computer vision in a docker image.

Easily get started with just one command.

Key Features

  • Docker container exposing a REST API
  • Available in CPU and GPU-accelerated versions
  • Able to interface with any language through HTTP calls
  • Deployable on premises or on your cloud infrastructure

Current Capabilities

Face recognition
(1:1 and 1:N face matching)
Face extraction and
landmark detection
Gender detection
Spoof detection
docker run -e "TOKEN=$YOUR_TOKEN" -v "$(pwd)"/collections:/collections -p 8080:8080 -it trueface/visionbox:latest

Supported Architectures

  • CPU: 1 Core
  • Memory: 2GB, 8GB recommended
  • Nvidia GPU optional but recommended

Supported Operating Systems:

Minimum System Requirements:

  • CPU: 1 Core
  • Memory: 2GB, 8GB recommended
  • Nvidia GPU optional but recommended

Supported Operating Systems:

Benchmarks

Speed (CPU)
Speed (GPU)
Enrolling a person
470 ms
100 ms
Identifying a person (1:N) N = 1000 person collection
450 ms
90 ms
Extracting face information
120 ms
80 ms
Detecting spoof attempt
590 ms
330 ms
Comparing two face features
4 ms
3 ms

*All benchmarks were performed on Intel(R) Xeon(R) E-2176M  CPU @ 2.70GHz, 32GB Ram, Nvidia Quadro P2000

Trueface Visionbox Resources

SDK documentationSample appsPostman
Trueface SDKs
Trueface Visionbox
Trueface Aware

Tell us about your project.

Contact Us
© 214 Technologies Inc 2020
Trueface is designed and developed in Los Angeles, California