⏹️Crop Auto-Annotation
Crop Auto-Annotation - Interest Area Auto-Annotation
Last updated
Crop Auto-Annotation - Interest Area Auto-Annotation
Last updated
Crop Auto-Annotation, also known as Interest Area Auto-Annotation, in Unitlab Annotate automatically annotates your data in the selected areas using either built-in AI models or your own models (BYO). You simply identify an Interest Area and allow the AI models to annotate the data for you.
The Crop Auto-Annotation tool is available to all users and is free forever. Start using it today!
Accurate Annotation: Crop Auto-Annotation can annotate more accurately than Batch Auto-Annotation because you guide it on where to annotate, rather than annotating the entire image.
Flexibility: You can bring your own pre-trained models (BYO) by integrating them with Unitlab Annotate. Learn how to integrate your AI models with Unitlab Annotate on the Model Integration page.
Example: Person Detection, Pose Estimation, and Segmentation using Interest-Area Auto-Annotation
To use Crop Auto-Annotation, make sure you select an AI model for your project while creating it. As mentioned earlier, you can integrate your own pre-trained AI model instead of using Unitlab's built-in AI models. The process of creating the project follows the same steps as those guided in the "Setup a Project" section.
📌Setup a ProjectIt generates annotations based on predictions from either the built-in model or your integrated model.
Learn how to integrate your AI models with Unitlab Annotate on the Model Integration page.
Carefully review the generated annotations after using auto-annotation tools to ensure their accuracy in Unitlab Annotate. It's important to closely examine each annotation for correctness and consistency, making adjustments as needed to guarantee the highest quality of data for your projects.
If you select an AI model for your project during its creation, you see this tool in the Unitlab Annotate toolbar when you open an image for annotation.
Simply click on the crop auto-annotation tool and select an area you want to annotate, as shown in the screenshot above.
