DevMesh

AI Art Mirror

Dmitry Kanevsky

Dmitry Kanevsky (Intel)

Unknown

6 0
  • 0 Collaborators

The project intent is to demonstrate Deep Learning usage for the Art by applying neural network style transfer on live video stream. This will make personal experience of users with art. ...learn more

Project status: Under Development

RealSense™, Artificial Intelligence, Graphics and Media

Intel Technologies
OpenVINO, AI DevCloud / Xeon, Intel Opt ML/DL Framework, Movidius NCS

Code Samples [1]Links [1]

Overview / Usage

Deep learning becomes more and more popular in Art word.
It allows to bring the Art to the life with personalized experience.

Methodology / Approach

  1. Train neural network on Intel® AI DevCloud.
    Utilizes the hardware and software optimized and high performance machines from Intel
  • Currently can train only on the laptop

  1. Optimize the models with Intel's OpenVino

  2. Run the live stream application with RealSense camera.
    The intent is to use depth stream to replace background with the original style image.

  3. Replace the high performace laptop with UP board and Movidius stick.
    *UP board is small size single board computer with Intel x86 64bit.

Technologies Used

Intel® AI DevCloud
Intel® OpenVino
Intel® RealSence laptop
fast-style-transfer
Nvidia® CUDA + CuDNN
TensorFlow GPU-enabled
OpenCV

*UP board and Movidius stick (in next step)

Repository

https://github.com/dimakan/ai-art-mirror

Comments (0)