Few-Shot Connectivity-Aware Text Line Segmentation in Historical Documents

TL;DR

This paper introduces a few-shot, connectivity-aware deep learning method for text line segmentation in historical documents, achieving state-of-the-art accuracy with minimal annotated data by using a lightweight architecture and specialized loss function.

Contribution

It demonstrates that small, simple models with a topology-aware loss outperform complex models in few-shot scenarios for historical document segmentation.

Findings

200% increase in Recognition Accuracy over previous methods

75% increase in Line Intersection over Union

Achieves or exceeds state-of-the-art F-Measure with only three annotated pages

Abstract

A foundational task for the digital analysis of documents is text line segmentation. However, automating this process with deep learning models is challenging because it requires large, annotated datasets that are often unavailable for historical documents. Additionally, the annotation process is a labor- and cost-intensive task that requires expert knowledge, which makes few-shot learning a promising direction for reducing data requirements. In this work, we demonstrate that small and simple architectures, coupled with a topology-aware loss function, are more accurate and data-efficient than more complex alternatives. We pair a lightweight UNet++ with a connectivity-aware loss, initially developed for neuron morphology, which explicitly penalizes structural errors like line fragmentation and unintended line merges. To increase our limited data, we train on small patches extracted from a mere three annotated pages per manuscript. Our methodology significantly improves upon the current state-of-the-art on the U-DIADS-TL dataset, with a 200% increase in Recognition Accuracy and a 75% increase in Line Intersection over Union. Our method also achieves an F-Measure score on par with or even exceeding that of the competition winner of the DIVA-HisDB baseline detection task, all while requiring only three annotated pages, exemplifying the efficacy of our approach. Our implementation is publicly available at: https://github.com/RafaelSterzinger/acpr_few_shot_hist.