loading1
loading2
alt

Project

Category

Research Methodology

Start

2021

Status

Completed

Acronym

Art_Bios_In_Me

ARTWORK BIOGRAPHIES AND INSTITUTIONAL MEMORY

Artworks in museum collections are managed and shaped by institutional policies and personal viewpoints. Oral History methodology can help to assess this socio-cultural perspective. However, the tools to facilitate this in conservation research are still lagging behind. The project ‘Artwork Biographies and Institutional Memory’ (Art_Bios_In_Me) is designed along two research lines addressing this problem. One explores obstacles and pitfalls in archiving and disclosing interviews to improve the museum workflow, while the other fosters technological advancements in transcribing this unique source material and disclosing it, preparing the material for digital humanities research.

The Kröller-Müller Museum (KMM) and the Cultural Heritage Agency of the Netherlands (RCE) will deliver case study material to assess the deposition infrastructure with Data Archiving and Networked Services (DANS), and carry out a feasibility study to adapt Automatic Speech Recognition (ASR) to the field of conservation with the Foundation for Open Speech Technology (FOST). Only with time-stamped transcripts, interviews become searchable – raising text-mining potential, cross-referencing, and other use of Artificial Intelligence (AI). The envisioned methodological and technological advancements in using Oral History methodology in conservation research will be of great advantage to technical art history, object diagnosis, conservation decision-making and analysis of cultural heritage at large.

 

Fig.1. Conservation language model generation workflow for ASR. New and existing interviews will be transcribed using Default ASR. The transcriptions will be manually corrected to get a 100% correct transcription. These corrected transcripts will be used to create and update the dedicated new Language Model (LM). This method is similar to the HoMed project to create a dedicated LM for medical interviews. ©Arjan van Hessen.