YOLOv8 AI Object Detection

Run object detection and output detected bounding boxes

Overview

The Cloneable platform supports deploying YOLOv8 models to edge devices running our platform. Train and implement your own models using the YOLOv8 framework. Easily implement the model into your apps in the app builder and the models will download to the device and run completely offline.

Usage

The component accepts an image as an input. You may send a single image, or images very quickly from a video. An example usage is to connect the output from the camera view component which outputs images frames at the full frame rate of the video.

Image input size

You can input images of any size and the model will adapt to it. The bounding box output includes the input image reference size in order to properly scale the bounding box back to the image.

Bounding box outputs

The component will output an array of bounding boxes for each image input. You may input these bounding boxes back into the camera component to visualize the bounding boxes on the camera feed. The camera will automatically scale and size the bounding boxes relative to the video on the device screen.

Bounding Box

Select your AI model

You must have a model uploaded. Once you have a model in Cloneable, select the model which this component will download and run. To select your model Click on the component to open the options menu and select your model from the drop down.

Offline usage

The model will automatically be synced offline to the devices that have access to this workflow and will be able to run offline.