Converting PDFs with Image Pages to Text

1min Snip

00:00

Play full episode

Summary

Transcript

Episode notes

PDFs can contain various types of content like embedded text, scanned pages, or a combination of both. When processing a PDF with image pages, they are treated as images and passed through an object detection layer to identify tables, headings, paragraphs, and narrative text. Tables within images are recognized as tables, and OCR steps are applied to extract content at the cell level. There are also experiments being conducted to convert actual images into a textual representation.

The majority of enterprise data exists in heterogenous formats such as HTML, PDF, PNG, and PowerPoint. However, large language models do best when trained with clean, curated data. This presents a major data cleaning challenge.

Unstructured is focused on extracting and transforming complex data to prepare it for vector databases and LLM frameworks.

Crag Wolfe is Head of Engineering and Matt Robinson is Head of Product at Unstructured. They join the podcast to talk about data cleaning in the LLM age.

Sean’s been an academic, startup founder, and Googler. He has published works covering a wide range of topics from information visualization to quantum computing. Currently, Sean is Head of Marketing and Developer Relations at Skyflow and host of the podcast Partially Redacted, a podcast about privacy and security engineering. You can connect with Sean on Twitter @seanfalconer .