Labelbox

Definitions

Definitions

Labeling

Adding information to data for the purpose of training machine learning algorithms. Also known as data annotation. For example, classifying if there is a car in an image.

Label

A label is a classification, a segmentation or a combination of classifications and segmentations that were created for a single data asset. For example, an image that was labeled with a car and a person, and the entire image was classified as being blurry.

Data Asset

A datarow is a single data asset. For example,this could be an image or text string that is imported into Labelbox to be labeled.

Project

A project is an organized set of training data to be labeled and managed by one or more collaborators.

Consensus

The agreement among multiple labelers and/or a model prediction on a training dataset.

Segmentation

A segment or a set of segments that collectively cover an image to classify object within an image for object detection. For example, a rectangle bounding box around a car in an image.

Classification

A Classification is an observation or multiple observations of a data asset.

Re-enqueue Labels

When labels are selected to go back to the labeling queue to be relabeled.