A glossary for definitions used in our products and documentation.

The nomenclature of any new platform can be overwhelming at first. We have tried to conform to industry naming standards wherever possible in a bid to make the user experience both intuitive and welcoming to new users and experienced practitioners alike. However, computer vision and annotation tooling are relatively recent fields, and as such many terms may be used interchangeably in both the literature and industry.

Do not hesitate to contact [email protected] for clarification on the features supported by Encord.



In the interests of brevity, we will refer to both image and video data as 'frames' in the definitions below.

AttributeAttributes can be nested into objects and classifications to provide more information on the label. For example, for the object 'cat,' an example attribute would be 'color.'
Benchmark functionThe function used to review tasks with benchmark QA. The benchmark function works by comparing all labels in the annotator submission of the benchmark task against the gold standard label set in the source project’s task.
Benchmark taskAn annotation task in a project with benchmark QA, which has a corresponding task in the ‘source project’ that contains gold standard labels.
Bounding boxA rectangle used to annotate a feature by drawing the bounds of the feature.
ClassificationA mutually-exclusive category applied to a frame.
Crosshair navigationA way to navigate in 3D. Clicking on a location in one slice will also change the associated views.
DatasetA collection of videos and/or images.
Data UnitA package of data that constitutes a single annotation task, e.g., a video, a single image, an image group, or a DICOM series.
FeatureAn object in a frame or a classification applied to a frame. These can be used to identify something in a frame (object: 'this thing is an apple') or to classify the frame itself (classification: 'this frame has apples').
Hanging protocolAn arrangement of views, e.g., Axial, sagittal, and coronal.
Hounsfield unitA linear transformation of the measured attenuation coefficient, e.g., air = -1000 HU, water = 0 HU.
Image groupA collection of images presented as one data unit. Grouping images in the image group functionality allows Encord's platform to support enhanced performance on playback and more automated labeling features. Also known as image sequences.
InstanceAlso known as an instance label in the platform, an instance is a unique instantiation of an ontology entity, which, depending on the data type, may contain many frame labels. For example, in a 100-frame video tracking three cars on a road, there are three instances of 'car' and up to 100 frame labels for each car.
Key pointA dot used to annotate a feature by specifying its location.
LabelSometimes denoted as a frame label in the platform, labels note relevant features in a frame and apply to a dataset used in model training. They are an annotation asserting which features in the desired ontology are true.
Label EditorThe UI for annotating data and managing labels.
Maximum intensity projection (MIP)A method for 3D data that projects all voxels to a plane.
Micro-modelA model specifically trained to label a dataset for training other models.
ModelA program with a set of functions and parameters that allow it to recognize features in datasets. Different models have different strengths and weaknesses.
Model trainingThe process of teaching a model an ontology. This is done by algorithmically changing model parameters until it can reliably recognize features that are labeled in a dataset.
Model inferenceThe process of using a trained model to predict the presence of features in new data.
ObjectSomething of interest in a frame. Defined by string together with an annotation. It can be used as part of an ontology to label entities of interest in a dataset used for model training. Examples include bounding boxes and polygons that have been applied to a frame.
Object detectionThe ability of a model to reliably recognize when a frame contains an object of interest. An application of model inference.
Object primitiveA unique object annotation type. Used to create templates of shapes (such as 3D cuboids and pose estimation skeletons) commonly used by your annotation team.
Object trackingThe ability of a model to reliably detect and track objects in a sequence of frames over time. An application of model inference.
OntologyA defined set of features and their relationships. This is what a model will be trained to apply to frames. Also known as a 'taxonomy.'
PolygonA polygonal shape used to annotate a feature by drawing the bounds of a feature.
PolylineA line composed of multiple segments.
ProjectA self-contained, collaborative environment for managing all productivity tasks associated with labeling and modeling one or more datasets.
QualityAn assessment of the accuracy of a set of labels.
Semantic segmentationThe application of labels to each pixel in a frame in order to classify segments of the frame as part of the same entity.
SliceA single image of a DICOM volume.
TaskAn action required as part of the labeling workflow.
Task ManagerThe UI for creating and managing tasks.
ViewA window displaying a specific viewing direction, e.g., coronal.
VolumeA set of images, also called slices or frames.
WindowingChanging the appearance of the image to highlight particular structures.