Multi-boundary prediction speeds up splitting scanned documents
Splitting Documents at Lower Cost: Multi-Split Boundary Decisions for LLM-Based Page Stream Segmentation
Artificial Intelligence
Summary
When scanned papers or PDFs arrive as long streams of pages, they need to be separated into individual documents before further use. The authors show that instead of detecting one page break at a time, a language model can predict multiple breaks in a group of pages, making the process faster. This works best if the group size is chosen carefully based on the type of documents and model used. The new method keeps accuracy high while reducing the number of requests to the model.
What this means in practice
- •For document processing teams: Reduce processing time in automated systems by predicting multiple document breaks per request for scanned or uploaded page streams.
- •For digital mailroom operators: Improve automation in sorting bundled attachments and scanned mail by using multi-boundary models to speed up document splitting without extra training.
Authors
Nikhil Reddy Pottanigari, Sepideh Kharaghani, Saverio Vadacchino, Alejandro Posada, Ying Zhang
Abstract
Scanned mail, uploaded PDFs, and consolidated attachments often arrive as page streams that must be split into individual documents before downstream classification, extraction, or routing. Zero-shot large language models can detect document boundaries without task-specific training, but standard Page Classification (PC) and Boundary Decision (BD) formulations resolve only one boundary per model call. We introduce Multi-Split Boundary Decision (MSBD), which predicts multiple boundaries within a page window in a single call, reducing the number of inference requests. We evaluate MSBD across multiple language models, document collections, input modalities, and window sizes. The results reveal a model- and corpus-dependent operating range in which MSBD preserves strong segmentation accuracy while substantially improving inference efficiency, followed by a sharp decline at larger windows. MSBD provided the strongest overall accuracy--efficiency trade-off, while large windows expose distinct over- and under-segmentation behavior across models. These findings show that multi-boundary prediction can make zero-shot page stream segmentation more efficient when the window size is selected for the target corpus.