Evaluating Hierarchy-Aware Deep Learning for the Recognition of Tironian Notes

2026-09-28

AI Summary

This study addresses the challenge of automatically recognizing Tironian shorthand symbols, which is hindered by their high visual similarity and severe data scarcity. To enhance recognition performance in low-resource scenarios, this work proposes exploiting the inherent hierarchical structure of these symbols. Methodologically, we introduce hierarchy-aware routing and a coarse-to-fine classification strategy, systematically comparing flat and hierarchical deep models—including HD-CNN, ViT, and ResNet—while incorporating few-shot domain adaptation for optimization. Experimental results demonstrate that HD-CNN achieves 45.43% accuracy without adaptation, whereas flat models reach 82.09% following domain adaptation. This research validates the effectiveness of structured priors in manuscript recognition, offering a novel paradigm for the digitization of low-resource ancient scripts.

Abstract

Tironian notes are generally regarded as the first Latin shorthand system and are notable for their large, fine-grained symbol inventory. Their high visual similarity and large class set make manual reading time-consuming, leaving manuscripts that contain Tironian notes inaccessible to many researchers. Automatic recognition is also challenging because models must distinguish subtle differences in stroke shape and sign structure while realistic training data remain scarce. However, standard flat classifiers do not explicitly use visual or structural relations between related signs. This paper investigates whether structural relationships between Tironian notes can support automatic recognition. We use the Supertextus Notarum Tironianarum (SNT) by Martin Hellmann, which provides idealized sign forms and a hierarchical organization of Tironian notes. We compare flat ResNet18, ConvNeXt, Shifted Window Transformer (Swin), and Vision Transformer (ViT) classifiers with Hierarchical Deep Convolutional Neural Network (HD-CNN)-style coarse-to-fine models and hierarchy-aware routing models based on visual class cleaning and similarity-based re-clustering. The models are evaluated on handwritten samples and manuscript-domain samples from Vergilius Turonensis, both with and without limited few-shot adaptation to the manuscript domain. The results show that the relative performance of flat and hierarchical models depends on adaptation. On Vergilius Turonensis, HD-CNN achieves the best non-adapted Top-1 result with 45.43%, while flat classification reaches the best Top-1 result after few-shot adaptation with 82.09%. Overall, the results indicate that hierarchical structure can support Tironian note recognition, especially under non-adapted conditions.

Problem

• Tironian notes
• automatic recognition
• hierarchy-aware deep learning
• fine-grained classification
• few-shot adaptation

Innovation

• Tironian notes recognition
• hierarchy-aware deep learning
• HD-CNN
• few-shot adaptation
• visual class cleaning