Sama Glossary
Updated at April 23rd, 2025
This is a auto-generated Article of all your definitions within the glossary.
Glossary
This is a auto-generated Article of all your definitions within the glossary.
-
Active canvas
The specific canvas currently working on eg, in cases of sensor fusion, where the scene is presented in more than 1 output, and will be affected by the toolbar actions.
-
Add shape
Means to add fused shapes of the same object in the same or different canvas.
-
Annotated asset
An asset (lmage, video, or document) that has added information to provide context, explanations, or identifications.
-
Attached objects
Linked objects that have an association.
-
Data collection
Logical regroupement of assets based on specific criteria or parameters.
-
Embeddings
Vector encoding the asset, fused asset or a shape (shape + crop asset) representation Metadata that does not have sense on its own (embeddings). Need to be compared to other assets
-
Metadata
Assets characteristics existing in assets description.
-
Multiple canvas
Fused projects where the same scene is presented in more than 1 output.
-
Object
An item that exists in a scene and belongs to a class.
-
Object list
List of identified objects in the scene through shape creation.
-
Object taxonomy
Object is defined by a class, attributes and shapes types.
-
Output
All deliverables that are tracked, managed, and utilized within the platform as a result of the annotation process.
-
Parent objects
A single shape that represents objects with multiple parts within it.
-
Pre-computed metadata
Assets characteristics that can have sense on a unit assets (brightness, sharpness, etc.)
-
Prediction
The output generated by a model that has been trained on historical data.
-
Scene taxonomy
The categorization of scenes or layouts.
-
Single canvas
Non-fused projects where the scene is presented in only 1 output.
-
Split view
The layout is divided into two areas to show 3D and 2D side by side.
-
Taxonomy
A system of classification, a way to organize and categorize object and scene based on shared attributes.