SDS KoPub VDR: A Benchmark Dataset for Visual Document Retrieval in Korean Public Documents

TL;DR

This paper introduces SDS KoPub VDR, a large-scale benchmark dataset for visual document retrieval in Korean public documents, addressing language and structural complexity gaps in existing VDR benchmarks.

Contribution

It provides the first comprehensive Korean public document dataset with multimodal queries and human-verified annotations for evaluating VDR models.

Findings

Significant performance gaps in multimodal retrieval tasks.

State-of-the-art models struggle with cross-modal reasoning.

The dataset enables detailed evaluation of document understanding models.

Abstract

Existing benchmarks for visual document retrieval (VDR) largely overlook non-English languages and the structural complexity of official publications. To address this gap, we introduce SDS KoPub VDR, the first large-scale, public benchmark for retrieving and understanding Korean public documents. The benchmark is built upon 361 real-world documents, including 256 files under the KOGL Type 1 license and 105 from official legal portals, capturing complex visual elements like tables, charts, and multi-column layouts. To establish a reliable evaluation set, we constructed 600 query-page-answer triples. These were initially generated using multimodal models (e.g., GPT-4o) and subsequently underwent human verification to ensure factual accuracy and contextual relevance. The queries span six major public domains and are categorized by the reasoning modality required: text-based, visual-based, and cross-modal. We evaluate SDS KoPub VDR on two complementary tasks: (1) text-only retrieval and (2) multimodal retrieval, which leverages visual features alongside text. This dual-task evaluation reveals substantial performance gaps, particularly in multimodal scenarios requiring cross-modal reasoning, even for state-of-the-art models. As a foundational resource, SDS KoPub VDR enables rigorous and fine-grained evaluation and provides a roadmap for advancing multimodal AI in real-world document intelligence. The dataset is available at https://huggingface.co/datasets/SamsungSDS-Research/SDS-KoPub-VDR-Benchmark.