GLOSSARY

Active learning: A semi-supervised learning model where an oracle (who may be the user) plays an active role in the process. More precisely, unlike the classical model where data is known and imposed, in active learning it is the learning algorithm that requests information for specific data identified as being crucial to refining the decision.

Image classification: In computer vision, classification is the name used for a family of methods that allow images to be grouped into classes, the characteristic of which is that the images in the same class are similar while those in two different classes are dissimilar in the sense of the same vector of parameters characterizing each of the images to be classified.

Deep learning: Deep learning is a form of artificial intelligence derived from machine learning and based on an artificial neural network inspired by the human brain. This network is made up of dozens or even hundreds of "layers" of neurons, each of which receives and interprets information from the previous layer.

Detection: In computer vision, object detection (or object classification) is a method used to detect the presence of an object or class of objects in a digital image.

Artificial intelligence: "The set of theories and techniques used to create machines capable of simulating intelligence."

Machine Learning: Field of artificial intelligence research that uses mathematical and statistical approaches to give computers the ability to "learn" from data, i.e. to improve their capacity to perform tasks without being explicitly programmed for each one.

Region of interest: In computer vision and image processing, feature detection consists of highlighting areas of an image deemed to be "interesting" for analysis, i.e. those presenting noteworthy local properties. Depending on the method used, such areas may appear in the form of points, continuous curves (segmentation) or connected regions, rectangular or otherwise, that constitute the result of the detection.

Neural network: An Artificial Neural Network is a computer hardware and/or software system modelled on the neurons of the human brain. When such networks are organized into multiple layers, they constitute an example of Deep Learning-style technology.

SaaS: Software as a Service (SaaS) is a commercial software operating model in which the software is hosted on remote servers rather than installed on the user's machine.

Segmentation: an image processing operation that aims to group pixels together according to predefined criteria. The pixels are thereby grouped into regions, which constitute a tiling or partition of the image. For example, this may be used to separate foreground objects from the background.