Automatic tagging of medieval manuscript images using ai models
Automatic Iconographic Annotation of Medieval Illuminations: Comparing Generative Vision-Language Models and Supervised Discriminative Classifiers
Computer Vision and Pattern Recognition
Summary
Medieval manuscripts are being digitized in large numbers, creating a need to automatically label the images inside them with meaningful descriptions. The authors compare two types of AI models: one that generates labels from images and text together, and another that learns to recognize specific labels from examples. They find that the label-learning models are more precise for well-known concepts, while the generative models handle broader and more complex ideas better. Combining both types could improve future tools for exploring and organizing these ancient images.
What this means in practice
- •For digital archivists: Automatically tag medieval manuscript images with detailed iconographic concepts to aid archival search and organization.
- •For museum curators: Use hybrid AI systems combining two model types to assist in exploring and annotating medieval illumination collections.
Authors
Vincent Jolivet, Lucas Terriel, Chahan Vidal-Gor{è}ne
Abstract
The large-scale digitisation of medieval manuscripts creates a growing need for automated iconographic indexing. We evaluate two complementary paradigms for this task on 9,735 full-folio images annotated with 2,333 TIMEL concepts: generative Vision-Language Models (VLM) and supervised discriminative classifiers (SDC). The VLM experiments evaluate fine-tuning, structured JSON decoding, taxonomybased candidates generation, and deterministic reranking. The discriminative experiments evaluate flat, hierarchy-aware, and patch-attention heads over a frozen DINOv3 encoder. Our results highlight key functional differences: ML-Decoder (SDC) achieves the strongest overall retrieval performance (precision@1 of 0.696, MRR of 0.806), with loss calibration (Asymmetric Loss) serving as the primary performance lever. While SDCs provide superior precision and reliability within a closed vocabulary, VLMs offer strong semantic generalization and broad taxonomic coverage, particularly through structured JSON targets and verify-calibrate reranking (hierarchical F1 of 0.559). Furthermore, model behavior strongly depends on concept abstraction: SDCs excel at perceptually grounded descriptive concepts, whereas VLMs provide useful hypotheses for complex interpretative scenes. Finally, we find that enforcing explicit hierarchy within model architectures limits coverage compared to spatial feature aggregation. Overall, rather than competing approaches, SDCs and VLMs emerge as complementary components for future hybrid indexing systems combining structured retrieval with exploratory annotation assistance. As a methodological contribution, we define the soft F1 metric to provide a taxonomy-aware evaluation that credits semantically close predictions based on lowest common ancestors.