跨域文档目标检测：Benchmark套件及方法
Cross-Domain Document Object Detection: Benchmark Suite and Method
摘要：
将文档页面的图像分解成高级语义区域（例如，图、表格、段落），文档目标检测（DOD）是智能文档编辑和理解等下游任务的基础。DOD仍然是一个具有挑战性的问题，因为文档对象在布局，大小，纵横比，纹理等方面变化很大。在实践中出现了一个额外的挑战，因为大型标记的训练数据集仅适用于与目标域不同的域。我们研究跨域DOD，其目标是使用来自源域的标记数据和仅来自目标域的未标记数据来学习目标域的检测器。来自这两个领域的文档在布局、语言和体裁上可能会有很大的不同。我们建立了一个由不同类型的PDF文档数据集组成的基准套件，可用于跨域DOD模型的训练和评估。对于每个数据集，我们提供页面图像、限界框注释、PDF文件以及从PDF文件中提取的渲染层。此外，我们提出了一种新的跨域DOD模型，该模型建立在标准检测模型的基础上，并通过合并三个新的对齐模块来解决域偏移问题：特征金字塔对齐（FPA）模块，区域对齐（RA）模块和渲染层对齐（RLA）模块。在基准测试套件上的大量实验证实了所提出的三个模块的有效性，所提出的方法显着优于基线方法。项目页面位于\url{https://github.com/kailigo/cddod}。
英文摘要：
Decomposing images of document pages into high-level semantic regions (e.g., figures, tables, paragraphs), document object detection (DOD) is fundamental for downstream tasks like intelligent document editing and understanding. DOD remains a challenging problem as document objects vary significantly in layout, size, aspect ratio, texture, etc. An additional challenge arises in practice because large labeled training datasets are only available for domains that differ from the target domain. We investigate cross-domain DOD, where the goal is to learn a detector for the target domain using labeled data from the source domain and only unlabeled data from the target domain. Documents from the two domains may vary significantly in layout, language, and genre. We establish a benchmark suite consisting of different types of PDF document datasets that can be utilized for cross-domain DOD model training and evaluation. For each dataset, we provide the page images, bounding box annotations, PDF files, and the rendering layers extracted from the PDF files. Moreover, we propose a novel cross-domain DOD model which builds upon the standard detection model and addresses domain shifts by incorporating three novel alignment modules: Feature Pyramid Alignment (FPA) module, Region Alignment (RA) module and Rendering Layer alignment (RLA) module. Extensive experiments on the benchmark suite substantiate the efficacy of the three proposed modules and the proposed method significantly outperforms the baseline methods. The project page is at \url{https://github.com/kailigo/cddod}.
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。
