metadata

pipeline_tag: object-detection

YOLOv8n validated for Unity Sentis (Version 1.4.0-pre.3*)

This is Yolov8n object detection example code for unity sentis, not real-time inference, for instance image.

** this is almost same code with ref, but only for image. **

YOLOv8n is a real-time multi-object recognition model confirmed to run in Unity 2023.

How to Use

First get the package com.unity.sentis from the package manager. You will also need the Unity UI package.

Create a new scene in Unity 2023 or higher (include in Unity 6000).
Install com.unity.sentis version 1.4.0-pre.3 from the package manager
Add the c# script to the Main Camera.
Create a Raw Image in the scene and link it as the displayImage
Drag the yolov8n.sentis or yolov8n.onnx file into the model asset field
Drag the classes.txt on to the labelAssets field
Place a *.png image file in the Assets folder and drag onto Input Texture field
Set the fields for the bounding box texture sprite (you can create your own one using a transparent texture or use an inbuilt one) and the font

If working correctly you should see something like this:

The NMS selection will be improved in later versions of Sentis. Currently uses singular-class approach.

Unity Sentis is the inference engine that runs in Unity 3D. More information can be found at here

The YOLO models use the GPLv3 license.