Perceptive Enhanced Realities of Colored collEctions through AI and Virtual Experiences

Project Information

PERCEIVE

Grant agreement ID: 101061157

DOI

10.3030/101061157

Project closed

EC signature date 9 December 2022

Start date 1 February 2023

End date 31 January 2026

Funded under

• Culture, creativity and inclusive society

Total cost

€ 3 785 358,75

EU contribution

€ 3 785 358,75

Coordinated by

CONSIGLIO NAZIONALE DELLE RICERCHE
Italy

Periodic Reporting for period 2 - PERCEIVE (Perceptive Enhanced Realities of Colored collEctions through AI and Virtual Experiences)

Reporting period: 2024-05-01 to 2026-01-31

Summary of the context and overall objectives of the project

Coloured cultural heritage collections—including ancient polychrome sculpture, paintings, textiles, historical photographs and born-digital artworks—are among the most fragile and complex categories of heritage to preserve, study and communicate. Their colours often fade, darken, yellow or disappear over time, making scientific interpretation difficult and limiting both conservation strategies and public understanding. At the same time, museums and cultural institutions increasingly need reliable digital methods to document, reconstruct, visualise and communicate colour change, while ensuring accessibility, authenticity, reuse and long-term sustainability of data and results.
In this context, PERCEIVE aimed to create a new way to preserve, perceive, curate, exhibit, understand and access coloured cultural heritage collections and digital artworks through an integrated combination of heritage science, AI, image-based rendering, interoperable digital infrastructures and hybrid visitor experiences.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

The activities performed and main achievements obtained between M1 and M15 include: the submission of 12 deliverables; definition of a scientific and theoretical base for the entire project; Development of the Concept for the Open Space Museum; set-up of five Working Groups; Set-up the common SharePoint; Development of a NextCloud Federated Cloud; Started an extensive Data acquisition campaign; Designed the PERCEIVE colour knowledge repository; Defined the Requirements for WP4 and WP5; Started to develop AI and IBR core, platform, tools and services; Carried out Co-Design activities focused on the development of the prototypes; Created a visual identity and logo; Developed the PERCEIVE website and Newslettet (www.perceive-horizon.eu); Co-organised a scientific conference (https://www.gch2023.eu/); Prepared an open access book on the PERCEIVE theory and technologies (‘Chromatic Visions’); Initiated networking activities with the Digital Cultural Heritage Cluster network.
This final reporting period (M15–M36, May 2024–January 2026) marks PERCEIVE’s transition from foundational setup and early prototyping to the delivery, validation, and public deployment of its main results:
(i) consolidated scientific protocols and datasets for coloured heritage scenarios;
(ii) AI/IBR methods and training pipelines.
(iii) an operational platform with reusable tools and services; and
(iv) experience prototypes and evaluation frameworks that translate research outputs into visitor-facing and professional applications.
Across M15–M36 the consortium consolidated an end-to-end workflow for “coloured collections” that goes from acquisition and knowledge structuring to material/appearance modelling and AI-IBR-enabled processing, to access through platform services, and finally to interpretation, mediation, and assessment through prototypes and user studies.
Key advancements include:
- Data governance, quality and compliance were finalised and integrated into consortium practice (final DMP; final quality/risk framework; policy recommendations) [D1.2 D1.4]
- Scientific workflows and repository infrastructure were completed: protocols and datasets were formalised and the Colour Knowledge Repository extended with interoperability mechanisms and API; the “dynamic data workflow” approach was consolidated to manage heterogeneous scenario data end-to-end [D3.3].
- Material appearance and colour-change modelling matured into documented methods covering physically based, data-driven, and hybrid approaches across the PERCEIVE scenarios [D3.2].
- AI/IBR core and training / inference framework were delivered as reusable architectures, including training data strategies and consumption patterns compatible with platform services [D4.2].
- The PERCEIVE Platform (final) was documented and validated, providing a production-oriented service ecosystem and an integrated toolset (e.g. analysis, reconstruction, stylisation, light-damage estimation, visualisation) [D5.2].
- Experience prototypes reached deployment-ready maturity (Caring, Authenticity, Participation/Open Space Museum), supported by implementation templates and replication-oriented documentation; the Design Toolbox was finalised as a practitioner resource [D6.1 D6.2]
- Main exhibitions have been designed and set-up in public events and museums as a bridge with the scientific and citizens community and a place for evaluation activities (DigitalHeritage 2025 “Perceive Isis’ Colours” and Perceive tools & services, Munch “Tiny Conservator” exhibition, Trondheim Museum “Fragile Colours” exhibition [D8.3];
- A structured evaluation strategy and assessment campaigns were completed, providing both methodological framing and field-based results for tools and prototypes [D7.2].
- Dissemination, exploitation and clustering activities culminated in major public/scientific moments (mainly the world congress Digital Heritage 2025), publications (as the two volumes Chromatic Visions published with Springer in open access), and demonstrator showcases [D8.3 Chromatic Visions, Annex 1]
- Training activities in the form of 13 online webinars, “PERCEIVE talks”, published in the YouTube channels [D2.3]

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

[Domain: Color Imaging]
Color is critical in any digitization, preservation or restoration project in cultural heritage; PERCEIVE has advanced the state of the art by integrating colour imaging with structured acquisition protocols, interoperable data workflows and reusable digital infrastructures for heterogeneous heritage scenarios.

[Domain: Color Analysis and Color Reconstructions and Prediction]
Color reconstruction of residual polychromies is based on the analysis of remaining traces of pigments; PERCEIVE has advanced the state of the art by combining analytical investigation, material appearance modelling, mock-up experimentation and AI-supported reconstruction workflows into a shared and transferable methodology.

[Domain: Color reconstruction and color changes prediction in paintings and works on paper]
Darkening, fading and yellowing often accompanied by flaking, crumbling and chalking of the paint affect thousands of iconic paintings and works on paper, with dramatic threats for their understanding, preservation and management; PERCEIVE has advanced the state of the art by developing integrated workflows for analysis, simulation and prediction of colour change to support conservation decisions and public communication.

[Domain: Color reconstruction of textiles]
Collections of historical textiles suffer colour degradation that limits both preservation and public access; PERCEIVE has advanced the state of the art by developing a methodology that combines non-invasive acquisition, mock-up creation, appearance modelling and AI/IBR-based simulation to accelerate the analysis, prediction and visual reconstruction of fading textiles.

[Domain: Color reconstruction and animation of photographs]
Old photographs can rapidly lose or distort their original colours, especially in processes such as Autochromes and lenticular films; PERCEIVE has advanced the state of the art by developing AI-based tools for colour reconstruction, defect correction and historical colour image generation across fragile photographic formats.

[Domain: AI and Machine Learning methods to color and lighting reconstruction]
Artificial Intelligence and Machine Learning can give cultural institutions, scientists and creative industries a major advantage in automating complex manual processes; PERCEIVE has advanced the state of the art by delivering reusable neural architectures, synthetic data strategies and service-oriented deployment frameworks for colour, appearance and lighting reconstruction tasks.

[Domain: Exhibition Design and Virtual Experiences]
Museums are increasingly innovating to engage broader publics; PERCEIVE has advanced the state of the art by introducing theory-driven, evaluation-based design frameworks and deployable hybrid prototypes centred on care, authenticity and participation in the communication of coloured heritage.

Project Logo