DKDS: A Benchmark Dataset of Degraded Kuzushiji Documents with Seals for Detection and Binarization

2025-11-12

AI Summary

Existing OCR methods suffer significant performance degradation on degraded Kuzushiji (classical Japanese cursive script) documents due to severe text degradation and overlapping red seals—yet no benchmark dataset supports joint evaluation of text and seal processing. Method: We introduce the first mixed-annotation dataset for degraded Kuzushiji manuscripts, featuring pixel-level text and seal masks across diverse degradation types and realistic seal occlusion scenarios. We propose a dual-task evaluation framework—simultaneous detection and binarization—and systematically assess YOLO-based models for joint text/seal detection, comparing traditional binarization, K-means clustering, and GAN-enhanced segmentation. Contribution/Results: Our experiments expose critical limitations of current methods under strong degradation, establishing reproducible baselines and quantitative metrics. This work fills two longstanding gaps: (1) a standardized, degradation-aware benchmark for historical document OCR, and (2) a unified evaluation protocol integrating detection and binarization for complex paleographic materials.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Kuzushiji, a pre-modern Japanese cursive script, can currently be read and understood by only a few thousand trained experts in Japan. With the rapid development of deep learning, researchers have begun applying Optical Character Recognition (OCR) techniques to transcribe Kuzushiji into modern Japanese. Although existing OCR methods perform well on clean pre-modern Japanese documents written in Kuzushiji, they often fail to consider various types of noise, such as document degradation and seals, which significantly affect recognition accuracy. To the best of our knowledge, no existing dataset specifically addresses these challenges. To address this gap, we introduce the Degraded Kuzushiji Documents with Seals (DKDS) dataset as a new benchmark for related tasks. We describe the dataset construction process, which required the assistance of a trained Kuzushiji expert, and define two benchmark tracks: (1) text and seal detection and (2) document binarization. For the text and seal detection track, we provide baseline results using multiple versions of the You Only Look Once (YOLO) models for detecting Kuzushiji characters and seals. For the document binarization track, we present baseline results from traditional binarization algorithms, traditional algorithms combined with K-means clustering, and Generative Adversarial Network (GAN)-based methods. The DKDS dataset and the implementation code for baseline methods are available at https://ruiyangju.github.io/DKDS.

Problem

• Addresses degraded Kuzushiji document recognition challenges
• Introduces dataset for text/seal detection and binarization tasks
• Provides baselines for handling document degradation and seals

Innovation

• Introduces DKDS dataset for degraded document analysis
• Uses YOLO models for text and seal detection
• Applies GAN-based methods for document binarization