Create Project - UI - Images and Videos

You’ve got your data into the Encord, so now you need to create a Project.

Projects consist of Ontologies, Workflows, Datasets, and Collaborators.

We’ve already got the E2E Fruit Dataset, so we’re going to make the Ontology. We’ll use the default Workflow and add some Collaborators.

1. Create Ontology

   Our Ontology E2E Fruit Ontology is going to have a number of object label types and classifications.

   

   This Ontology is used for all Project creation types, from Basic to SOTA.

   Class name	Label Type	Attributes
   Blueberries	Polygon	Radio button: Type? Bluegold Duke Blueray
   Cherries	Bounding box	Radio button: Type? Bing King Rainier
   Apples	Bitmask	Radio button: Types? Sugar Bee Granny Smith Honey Crisp
   Bananas	Polygon	Radio button: Type? Cavendish Red Blue Java
   Kiwi	Bounding Box	Radio button: Type? Golden Green Kiwi Berry
   Persimmons	Bounding Box	Radio button: Type? Hachiya Fuyu Rojo Brilliante
   Strawberries	Bounding Box	Radio button: Type? Albion Redcoat Sweet Kiss
   Ripe?	Classification	Radio button: Yes No
   Type of fruit?	Classification	Radio button: Blueberries Cherries Apples Bananas Kiwi Persimmons Strawberries Other > What type? - Text box

   Start Ontology Creation

   

   Add Objects

   

   Add Classifications

   

   Save Ontology

   

2. Create Project

   This step brings the Dataset and Ontology you created together with Collaborators and the Workflow.

   We’ll use the default Standard workflow for the project. This is what the Standard workflow looks like:

   

   To create the Project:

   1. Click Projects > New annotation project.

   2. Add the following:

      • Project name: E2E - Fruit - Project I
      • Ontology: E2E Fruit Ontology
      • Dataset: E2E Fruit Dataset
      • Collaborators: Add your Annotators (label data), Reviewers (review labels), Team Managers (manage Annotators and Reviewers), and Project Admins (manage Project and export labels).
      • Workflow: We’ll use the default Standard workflow for the Project.

   

3. Label Data

   With your Project creation complete, you and your team are now ready to label your data.

   

   Manually Label

   Manually annotate and review your data and then export your labels.

   

   Use SAM 2

   Use SAM 2 and SAM 2 tracking to annotate your data, review using humans, and then export your labels using the UI or the SDK.