What is Supervisely: the free video course.

No single labeling interface is best for every task: it should fit the task and the skills of the person labeling, and use AI wherever it can. The demos label images, videos, LiDAR point clouds and DICOM volumes, deploy AI models and use them to label each type of data, and customize the image and video labeling interfaces.

• Image labeling
• Video labeling
• 3D point cloud labeling
• DICOM labeling
• AI-assisted labeling
• Custom labeling UI

How annotators, reviewers, domain experts and data scientists work together in a transparent, systematic way. The demos organize data with teams, users and roles, follow a team as it labels a set of images, and check the quality of the labels.

• User collaboration
• Quality assurance

Why Supervisely is a platform rather than a single product: apps extend and customize it for new tasks. The slides cover how apps are built, and the demos turn a raw video into labeled images, tour the ecosystem of apps, customize the code of an existing app and use the Python SDK in a Jupyter notebook.

• Ecosystem of apps
• Platform overview

State-of-the-art models are scattered across GitHub repositories; Supervisely brings them together so you can use them through a GUI in day-to-day work. The chapter covers out-of-the-box models for pre-labeling and the Smart Tool, then custom models: a smart tool trained for crack segmentation, object detection with synthetic data and a detection model for seed recognition.

• Model training and serving
• Synthetic data

Enterprise-level features, support and complementary services, with the team making sure that customers keep solving their tasks. The chapter covers the Enterprise edition, how the team works with customers and the reasons to choose Supervisely.

• Enterprise edition