Contar hilos con inteligencia artificial: una IA pequeña, verificable y al servicio de la mirada experta

Counting threads in X-ray images of canvases reveals whether two or more paintings come from the same piece of fabric or whether a work was cut down. At the Universidad de Sevilla, together with the Technical Documentation Department of the Museo Nacional del Prado, we have developed deep learning models that estimate thread density across the whole radiograph. They build on Aracne, a signal processing tool that the Prado makes openly available, and overcome its limitations with irregular or very dense weaves. This is a discriminative AI, small and specialised, not a generative one. Its biases are known and stated: it learned almost exclusively from European plain-weave canvases, with 240 samples each labelled by a single expert. We argue for data protocols set by public institutions that protect their collections without giving up useful tools. The machine proposes and the specialist decides. We conclude that sustainable AI is proportionate; ethical AI, verifiable; and social AI, shared and developed together with the institutions.

Authors

Institutions

Publication Details

Journal
PH. Boletín del Instituto Andaluza del Patrimonio Histórico/PH/PH. Boletín del Instituto Andaluz del Patrimonio Histórico/Boletín informativo - Instituto Andaluz del Patrimonio Histórico
Published
2026-11-12
DOI
https://doi.org/10.33349/2027.120.6282
Primary Topic
Cultural Heritage Materials Analysis
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Contar hilos con inteligencia artificial: una IA pequeña, verificable y al servicio de la mirada experta

Juan José Murillo Fuentes
PH. Boletín del Instituto Andaluza del Patrimonio Histórico/PH/PH. Boletín del Instituto Andaluz del Patrimonio Histórico/Boletín informativo - Instituto Andaluz del Patrimonio Histórico
Cultural Heritage Materials Analysis
article

Contar hilos con inteligencia artificial: una IA pequeña, verificable y al servicio de la mirada experta

Juan José Murillo Fuentes
article en

Abstract

Counting threads in X-ray images of canvases reveals whether two or more paintings come from the same piece of fabric or whether a work was cut down. At the Universidad de Sevilla, together with the Technical Documentation Department of the Museo Nacional del Prado, we have developed deep learning models that estimate thread density across the whole radiograph. They build on Aracne, a signal processing tool that the Prado makes openly available, and overcome its limitations with irregular or very dense weaves. This is a discriminative AI, small and specialised, not a generative one. Its biases are known and stated: it learned almost exclusively from European plain-weave canvases, with 240 samples each labelled by a single expert. We argue for data protocols set by public institutions that protect their collections without giving up useful tools. The machine proposes and the specialist decides. We conclude that sustainable AI is proportionate; ethical AI, verifiable; and social AI, shared and developed together with the institutions.

PH. Boletín del Instituto Andaluza del Patrimonio Histórico/PH/PH. Boletín del Instituto Andaluz del Patrimonio Histórico/Boletín informativo - Instituto Andaluz del Patrimonio Histórico
Universidad de Sevilla (ES)
Cultural Heritage Materials Analysis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon