LIVA & CLIVA: a semantic processor for visual data and its graphical configurator

LIVA (LIst Visual Analyzer) offers semantic processing and, in particular, detection and classification, of visual data. CLIVA is a graphical user interface (GUI) used to configure LIVA.

Key elements


The LIVA (LIst Visual Analyzer) software suite, with its modular architecture, offers a range of semantic processing features—including detection and classification—for visual data.

LIVA was engineered for open data processing use cases and can manage the constraints—such as unconstrained sensors, large scales, and very wide semantic ranges—associated with these use cases.
Specialists in semantic, text and image analysis began developing LIVA in the 2010s as part of the DataScale project.

Image analysis methods such as classification, detection, and, in particular, intermediate semantic layers, are natively integrated.

Subsequent development work resulted in CLIVA, the configurator for the LIVA suite. CLIVA enables simple, shared, and incremental configuration and domain adaptation through a graphical interface and automated web search functions. With CLIVA, users can rapidly adapt LIVA to specific domains.

Non-specialist users can provide their own data and add subject-matter knowledge to the annotations. Reports estimating the quality of the desired results allow users to fine-tune LIVA’s settings.


LIVA produced excellent results for location, plant, and food recognition in several international academic research projects.
The plant and food recognition technologies were transferred to partners as part of the Groww and Nutriperso project.

The face recognition technologies was transferred to our partners such as FacePoint.


The CLIVA semi-automatic configuration interface allows users to add new identities (faces) for recognition


A multi-criteria image search interface connected to the LIVA-integrated search engine