Live!Check out the Label Studio 2024 recap post
Back to integrations

Integrate Ultralytics YOLO with Label Studio

Overview

The Label Studio ML backend integration for Ultralytics YOLOv8 and YOLO11 enables a broad range of real-time computer vision tasks, including advanced object detection, segmentation, classification, and video object tracking.

As the most flexible data labeling platform, Label Studio gives you a tremendous amount of control in how you integrate and interact with YOLOv8 and YOLO11, combining both automation and human-in-the-loop workflows. Users have the ability to choose any combination of labeling tasks and formats, making use pre-configured templates or customized interfaces.

This ML backend is straightforward to install and configure in Label Studio; realistically, you can be fully up and running within 10 minutes. Once it’s installed, you can immediately start leveraging your datasets with any of the YOLO tasks listed in the table below. For example, you can use YOLO to automatically label a dataset in Label Studio, you can use Label Studio to evaluate how YOLO performs on your own data, or you can create a training dataset to fine-tune the YOLO model.

Export annotations from Label Studio to the YOLO format to use the image annotations for training and fine-tuning models.

Use Cases

YOLO Task NameLS Control TagPrediction SupportedLS Import SupportedLS Export Supported
Object Detection<RectangleLabels>YOLO, COCOYOLO, COCO
Oriented Bounding Boxes (OBB)<RectangleLabels model_obb="true">YOLOYOLO
Image Instance Segmentation: Polygons<PolygonLabels>COCOYOLO, COCO
Image Semantic Segmentation: Masks<BrushLabels>NativeNative
Image Classification<Choices>NativeNative
Pose Detection<KeyPoints>NativeNative
Video Object Tracking<VideoRectangle>NativeNative
Video Temporal Classification<TimelineLabels>NativeNative

Related Integrations

Grounding DINO

Text-Driven Object Detection Model

Nvidia NeMo

Automated audio transcription

Scikit Learn

Machine learning toolkit

LangChain

Evaluate LLM Output Quality