Exploring the "Great Unseen" in Medieval Manuscripts: Instance-Level Labeling of Legacy Image Collections with Zero-Shot Models

TL;DR

This paper proposes a holistic approach to medieval manuscript analysis by using advanced segmentation and multimodal techniques to better understand and label legacy image collections, enhancing training data for computer vision.

Contribution

It introduces a novel method for instance-level labeling of medieval manuscripts using zero-shot models, improving data richness for visual content analysis.

Findings

Enhanced segmentation of manuscript pages

Improved multimodal content description

Richer training data for computer vision

Abstract

We aim to theorize the medieval manuscript page and its contents more holistically, using state-of-the-art techniques to segment and describe the entire manuscript folio, for the purpose of creating richer training data for computer vision techniques, namely instance segmentation, and multimodal models for medieval-specific visual content.