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

Core entities

Image

A file that contains a visual representation of something. for example, a GIF or PNG file. An image can be an artifact, such as a photograph or other two-dimensional picture, that resembles a subject.

Video
A file that contains the recording, reproducing, or broadcasting of moving visual images.

Video Frame

An individual image extracted from a video sequence. Each frame represents a single moment in time and can be analyzed or processed independently.

Object
A region of interest in an image, usually a region that contains an instance of a specific class.

Dataset

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

Exploration

Dataset exploration
The process of browsing, filtering and searching through the data in a Dataset, usually for data selection purposes.

Cluster
A set of data entities unified by a certain characteristic. Images are grouped by sophisticated similarity algorithm.

Semantic Search

A search technique that uses the meaning and context of terms rather than exact matches, enabling more intuitive and relevant results when querying datasets.

Similarity Search
The process of finding similar Images and/or Clusters for a given query Image or Cluster

Similarity Vertex
The item/s (Image, Cluster or Object) that user chooses to see its similar Images

Similarity results
Images and Clusters that are visually similar to a chosen Vertex

Selected item/s
The item/s (Image, Cluster or Object) that user selects in order to see details about or perform an action on

Filtering
process of reduction of a set of items by applying conditions (query expression) to items metadata properties. Set of items can be anything: a cluster, entire dataset, similar images

Quality 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.

Issue Type
A specific type of a problem that during the analysis process takes one or more images in and generates zero or more issue instances.

Issue
A concrete instance of an issue type, associated with one or more images (e.g. train/test leakage, cluster of duplicates, data drift).

Duplicates

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

Mislabel

An incorrect or inconsistent label applied to a dataset item. Identifying and correcting mislabels is crucial for maintaining dataset accuracy and reliability.

Outliers

Data points that deviate significantly from the rest of the dataset. Visual Layer enables users to identify and analyze outliers to improve data quality and gain insights.

Metadata

Metadata

Information associated with each visual data item, such as file name, dimensions, creation date, or user-defined attributes. Metadata plays a key role in filtering and organizing datasets.

Enrichment

The process of adding additional metadata or attributes to dataset items to enhance their value and utility. This can include annotations, tags, or external data sources.

Data organization

Tags

User-defined labels applied to dataset items to categorize, group, or annotate specific data points. Tags enhance searchability and organization.

Saved Views

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