NovaLAD用CPU快速解析文档，支持生成式AI与数据智能应用。
NovaLAD: A Fast, CPU-Optimized Document Extraction Pipeline for Generative AI and Data Intelligence
- 双并行YOLO模型同步检测内容与版面结构，提升效率。
- 在DP-Bench上达到96.49% TEDS与98.51% NID，优于商用开源方案。
- 无需GPU、支持多格式输出，适合部署在低资源环境。
文档提取是检索增强生成（RAG）、知识库及下游生成式AI的关键前置步骤，将非结构化文档如PDF和扫描件转化为结构化文本与布局感知表示。本文提出NovaLAD，一个集成双并行YOLO检测模型（元素检测与版面检测）的完整文档解析系统，结合规则分组与可选视觉-语言增强。输入页面图像后，两模型同时运行：元素模型识别标题、正文、表格、图像等语义内容；版面模型识别layout_box、column_group、multi_column等结构区域。关键设计为先通过ViT图像分类器判断图像相关性，仅保留有用图像送入视觉大模型进行标题、摘要与结构信息提取，降低噪声与成本。NovaLAD专为速度优化：可在CPU上运行，采用并行处理检测、分类、OCR与转换，输出结构化JSON、Markdown、RAG就绪文本及知识图谱。在DP-Bench基准（upstage/dp-bench）上取得96.49% TEDS与98.51% NID，优于商业与开源解析器。本文详述数据提取流程、架构设计与实现方式，展示如何在无GPU情况下兼顾精度与实用性。
原文摘要 · Abstract (English)
Document extraction is an important step before retrieval-augmented generation (RAG), knowledge bases, and downstream generative AI can work. It turns unstructured documents like PDFs and scans into structured text and layout-aware representations. We introduce NovaLAD, a comprehensive document parsing system that integrates two concurrent YOLO object detection models - element detection and layout detection - with rule-based grouping and optional vision-language enhancement. When a page image is sent in, the first thing that happens is that it goes through both models at the same time. The element model finds semantic content like the title, header, text, table, image, and so on, and the layout model finds structural regions like layout_box, column_group, multi_column, row_group, and so on. A key design decision is to first send an image or figure through an image classifier (ViT) that decides whether it is relevant or not. Only useful images are then submitted to the Vision LLM for title, summary, and structured information, which cuts down on noise and costs. NovaLAD is built for speed: it works on CPU, employs parallel execution for detection, classification, OCR, and conversion, and generates several forms, including structured JSON, Markdown, RAG-ready texts, and knowledge graphs. We test on the DP-Bench benchmark (upstage/dp-bench) and get 96.49% TEDS and 98.51% NID, which is better than both commercial and open-source parsers. This paper explains how to extract data, how the architecture works, how data flows, and how to make NovaLAD both accurate and usable without needing a GPU.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。