DocSpiral: A Platform for Integrated Assistive Document Annotation through Human-in-the-Spiral

TL;DR

DocSpiral tackles the challenge of extracting structured data from domain-specific, image-based documents by proposing a Human-in-the-Spiral workflow that iteratively couples human validation with model-driven annotation. The platform unifies document format normalization (Anything2PDF), a comprehensive annotation interface, and an AI/ML model enhancement layer, enabling progressive improvements across multiple cycles. Key contributions include a full-featured end-to-end annotation system, an assisted spiral improvement framework, and an open, deployable solution that lowers barriers to AI adoption in document-intensive domains. Empirically, the approach reduces annotation time by at least $41\%$ and shows consistent gains across three iterations, supporting scalable development of domain-specific AI for geoscience and healthcare applications.

Abstract

Acquiring structured data from domain-specific, image-based documents such as scanned reports is crucial for many downstream tasks but remains challenging due to document variability. Many of these documents exist as images rather than as machine-readable text, which requires human annotation to train automated extraction systems. We present DocSpiral, the first Human-in-the-Spiral assistive document annotation platform, designed to address the challenge of extracting structured information from domain-specific, image-based document collections. Our spiral design establishes an iterative cycle in which human annotations train models that progressively require less manual intervention. DocSpiral integrates document format normalization, comprehensive annotation interfaces, evaluation metrics dashboard, and API endpoints for the development of AI / ML models into a unified workflow. Experiments demonstrate that our framework reduces annotation time by at least 41\% while showing consistent performance gains across three iterations during model training. By making this annotation platform freely accessible, we aim to lower barriers to AI/ML models development in document processing, facilitating the adoption of large language models in image-based, document-intensive fields such as geoscience and healthcare. The system is freely available at: https://app.ai4wa.com. The demonstration video is available: https://app.ai4wa.com/docs/docspiral/demo.

Figures (8)

• Figure 1: Our DocSpiral framework converts documents to PDF and processes them through iterative cycles where human verification creates annotations that improve AI/ML models, reducing effort and enhancing performance within each iteration.
• Figure 2: System Architecture Overview for DocSpiral
• Figure 3: Documents upload and management interface, users can drag and drop allowed format documents or zipped files. Files will be uploaded to S3 bucket, and downstream tasks will be triggered.
• Figure 4: Layout annotation interface: user can click, add or remove bounding boxes from PDF Viewer, and assign layout labels (middle), or customize a domain specific hierarchical layout schema (right).
• Figure 5: OCR verification and annotation interface