Generating Synthetic Data for Image Segmentation with Unity and PyTorch/fastai скачать в хорошем качестве

Generating Synthetic Data for Image Segmentation with Unity and PyTorch/fastai

This tutorial 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. Check out the blog post for more details: https://blog.stratospark.com/generati... Here is the GitHub repo: https://github.com/stratospark/UnityI... Here's a link to the original Unity ML-ImageSynthesis repo: https://bitbucket.org/Unity-Technolog... Some of the topics covered in this video include: Downloading the ML-ImageSynthesis code and exploring its functionality Basic components of the Unity Editor GUI and how to customize it Organizing a Unity project with appropriate folders Creating new Scenes Creating new GameObjects, specifically 3d solids like Cubes, Spheres, and Cylinders Manipulating objects in the Scene view, including positioning, rotating, and scaling with mouse and keyboard Using the Inspector pane to modify public properties of GameObject components The difference between Scene and Game views and how to adjust the Camera Creating user-defined Layers to act as object categories Creating and modifying Prefabs Creating a custom C# component script Exposing script variables to the Editor GUI Visually linking components to each other in the GUI Adding custom camera display resolutions Modifying object appearance with custom Materials Enabling the Physics Engine and how RigidBodies work The Start/Update methods of MonoBehavior subclasses Randomizing the position/rotation/scale/color of a GameObject through code Inspecting memory usage with the Profiler Creating an Object Pooling system to reuse objects and cap memory usage Modifying ImageSynthesis code to output only a specific annotation image Creating fields to specify # of training and validation images in the Editor Modifying the layer colors to conform to the grayscale RGB values the image segmentation network requires Using the fastai Datablock API to load data produced by our Unity simulation Sorry about the video and audio sync problems! I'm new at this type of video production and hope to improve for future tutorials.