Skip to content

Use Unity to generate synthetic images for deep learning image segmentation in PyTorch and fastai

License

Notifications You must be signed in to change notification settings

stratospark/UnityImageSynthesisTutorial1

Repository files navigation

Unity Image Synthesis

This project will help you get up to speed with generating synthetic training images in Unity. You don't need any experience with Unity, but experience with Python and the fastai library/course is recommended. By the end of the tutorial, you will have trained an image segmentation network that can recognize different 3d solids.

Read more details on my blog

Follow along with the video tutorial

Getting Started

  1. Clone the repo
  2. Open the project with Unity / Unity Hub (2018.3.2 recommended)
  3. Open up the "Solids" scene
  4. Create a "captures/train" and "captures/val" folders
  5. Open up the SceneControl object and enable Save/grayscale if you want to do fastai training. Otherwise, leave those disabled and just see the annotations within Unity.
  6. Press the Run button
  7. Create a secondary Game window and change the Display to Display 3 for layer based category annotations.

About

Use Unity to generate synthetic images for deep learning image segmentation in PyTorch and fastai

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published