DeepMine-Mamba: Mitigating Information Dilution in Mamba-Based State Space Models for Document Image Binarization

2026-06-07

AI Summary

This work addresses the challenge of information dilution in existing Mamba-based state space models for document image binarization, where long-range modeling tends to weaken critical foreground cues such as faint ink or broken strokes. To mitigate this issue, we introduce DeepMine-Mamba—the first framework to adapt Mamba to this task—featuring a novel Anti-Dilution Gate mechanism. This gate selectively enhances stroke-sensitive local responses and suppresses background interference by estimating feature propagation dynamics. Evaluated on the DIBCO and H-DIBCO benchmarks, our method achieves competitive average F-measure and Fps scores. Ablation studies further demonstrate its effectiveness in significantly reducing perceptually salient binarization errors and substantially improving the preservation of fine and weak strokes.

Abstract

Document image binarization aims to separate foreground text from degraded backgrounds while preserving thin, broken, and low-contrast strokes. Although deep learning methods have improved binarization performance, most existing approaches rely on convolutional, transformer-based, or generative architectures, while Mamba-based state space models remain largely unexplored for this task. In this work, we investigate Mamba-based feature propagation and observe that direct state-space propagation may dilute weak foreground cues during long-range modeling, especially faint ink traces, fragmented characters, and boundary-sensitive stroke details. To address this problem, we propose DeepMine-Mamba, a Mamba-based binarization framework equipped with a novel Anti-Dilution Gate that estimates propagation-induced feature changes and selectively restores stroke-sensitive local responses while suppressing unnecessary background enhancement. Experiments on DIBCO/H-DIBCO benchmarks under a strict leave-one-year-out protocol show that DeepMine-Mamba achieves competitive overall performance, with strong average FM and Fps across benchmark years. Ablation results further demonstrate that the Anti-Dilution Gate improves stroke preservation and reduces perceptually significant binarization errors.

Problem

• Document Image Binarization
• Mamba
• State Space Models
• Information Dilution
• Stroke Preservation

Innovation

• Mamba
• state space model
• document image binarization
• Anti-Dilution Gate
• information dilution