基于Vision Mamba的自适应多尺度文档二值化
Adaptive Multi Scale Document Binarisation Using Vision Mamba
AI总结：
针对现有文档二值化混合模型存在的二次时间复杂度问题，提出基于Vision Mamba的架构，结合融入DoG特征的改进跳跃连接，实现长序列高效处理与高质量细节输出。
摘要：
增强并保持文档图像（尤其是历史文档图像）的可读性，对有效的文档图像分析至关重要。目前已有大量模型被提出用于该任务，包括基于卷积的架构、基于transformer的架构以及卷积-transformer混合架构。尽管混合模型解决了纯卷积或纯transformer方法的局限性，但它们常存在二次时间复杂度等问题。在本研究中，我们提出了一种基于Mamba的文档二值化架构，该架构通过线性缩放和内存使用优化，能够高效处理长序列。此外，受传统信号处理技术启发，我们通过引入高斯差分（DoG）特征对跳跃连接进行了新颖的改进。这些多尺度高频特征使模型能够生成高质量、细节丰富的输出结果。
英文摘要：
Enhancing and preserving the readability of document images, particularly historical ones, is crucial for effective document image analysis. Numerous models have been proposed for this task, including convolutional-based, transformer-based, and hybrid convolutional-transformer architectures. While hybrid models address the limitations of purely convolutional or transformer-based methods, they often suffer from issues like quadratic time complexity. In this work, we propose a Mamba-based architecture for document binarisation, which efficiently handles long sequences by scaling linearly and optimizing memory usage. Additionally, we introduce novel modifications to the skip connections by incorporating Difference of Gaussians (DoG) features, inspired by conventional signal processing techniques. These multiscale high-frequency features enable the model to produce high-quality, detailed outputs.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
