Deep learning vision for everyday devices

The world’s most lightweight deep-learning software made for embeddded CPU

Broadmann17 detects and locates objects in real time

Allowing applications such as augmented reality, robotics, home security, smart cities, and autonomous vehicles

Detection on smartphone – case study

To improve face detection in the smartphone camera without draining the battery, Brodmann17 implemented its deep learning on a single ARM A72 core. Achieving real-time with smallest face size of 25x25 and exceeding latest results in the popular benchmarks.

Broadmann17 detects and locates objects in real time

Allowing applications such as augmented reality, robotics, home security, smart cities, and autonomous vehicles

Speeding thing up

Brodmann17 Frames per second

  • Nvidia TitanX

    200+

  • Samsung ARTIK10 A-15

    25

  • Intel i7 7700HQ

    160

  • ARM A72 (single core)

    30+

  • Intel NUC i3 5010U

    60

  • Raspberry Pi 3

    4

  • Nvidia Jetson TX1

    40

Get in touch

We'd love to talk! Drop us your email and we'll get back to you ASAP!