This glossary provides definitions and explanations for key terms and entities relevant to Visual Layer. Use it as a quick reference to understand the platform’s core concepts and features.


Annotations

Descriptive metadata or markers added to images or objects, such as bounding boxes or class labels, to enhance dataset usability.

Bounding Box

A shape annotation drawn around an object in an image to define its position and size for object detection tasks.

Cluster

A set of data entities grouped by a common characteristic. Images are grouped by a sophisticated similarity algorithm.

Cluster View

A structured interface that visually presents groups of similar images, allowing users to explore related data efficiently.

Dataset

A collection of visual data, such as images or videos, that users upload to Visual Layer for organization, exploration, and analysis.

Dataset Analysis

A process in which the images in a given dataset are tested for possible issues from the known issue types and are grouped together for easy visualization. The result of the analysis is zero or more issues and zero or more Clusters.

Dataset Exploration

The process of browsing, filtering, and searching through dataset contents, typically for data selection or inspection.

Dataset Exploration View

The primary view for exploring and analyzing a dataset. It provides access to images, objects, clusters, metadata filters, and search tools.

Dataset Inventory

An overview of all available datasets in your Visual Layer environment.

Duplicates

Identical or near-identical images or videos within a dataset. Identifying and removing duplicates improves data quality and prevents redundancy.

Filtering

The process of reducing a set of items by applying conditions to metadata properties. Filters can be applied to the entire dataset, a cluster, or a similar images.

Enrichment

The process of adding metadata or attributes to dataset items to enhance their value. This includes annotations, tags, or external data sources.

Image

A visual file (e.g., PNG, JPG, GIF) representing an object or scene. It can be an artifact, such as a photograph or other two-dimensional picture, that resembles a subject.

Image View

Displays clusters of visually similar images across your dataset, enabling efficient visual browsing.

Issue

A concrete instance of an issue type, typically linked to one or more images (e.g., train/test leakage, duplicates, or data drift).

Issue Type

A predefined category of problem identified during dataset analysis. It generates issue instances based on specific patterns or anomalies.

Metadata

Information associated with each visual item, such as filename, dimensions, timestamps, and user-defined tags. Crucial for organizing and filtering datasets.

Mislabel

An incorrect or inconsistent label applied to a dataset item. Detecting and correcting mislabels ensures dataset accuracy.

Object

A region of interest within an image, usually representing an instance of a specific class (e.g., a face, vehicle, or product).

Object View

Displays clusters of annotated or enriched objects grouped by visual similarity. This view focuses on object-level insights rather than whole images.

Outliers

Data points that differ significantly from the rest of the dataset. Identifying outliers can improve data quality and uncover unique patterns.

Saved View(s)

Custom configurations of filters and sorting applied to a dataset. Saved Views allow users to revisit specific subsets of data and share them with collaborators.

Selected Item(s)

The image, object, or cluster a user selects to view in detail or perform an action on.

A search technique based on meaning and context rather than exact keyword matches, enabling more intuitive dataset queries.

Similarity Results

Images or clusters that are visually similar to a selected item (the “vertex”).

The process of finding visually similar images or clusters for a given query item.

Similarity Vertex

The reference item (image, object, or cluster) used as the starting point for similarity search.

User Tags

Custom labels applied by users to annotate or categorize dataset items. Tags support better organization, search, and collaboration.

Video

A media file containing a sequence of moving images. Videos are treated as datasets and can be analyzed frame-by-frame.

Video Frame

An individual still image extracted from a video sequence. Each frame can be analyzed or annotated independently.