Diagram Image Extraction Trust Report

Image extraction, thumbnail, and diagram-link confidence evidence.

Diagram Image Extraction Trust Report

Images are extracted only during the PDF-to-Markdown conversion stage. Downstream RFQX analysis consumes Markdown and converter sidecar metadata only.

• PDFs processed for image extraction: 10
• Images extracted: 121
• Thumbnails generated: 121
• Images ignored as decorative/small: 167
• Diagram artifacts detected: 87
• Diagram artifacts with image assets: 87
• Diagram artifacts without image assets: 0
• Embedded-image links: 41
• Page-snapshot links: 46
• High-confidence image links: 79
• Medium-confidence image links: 8
• Low-confidence image links: 0
• Pages with diagrams but no image: 0
• Converter limitations: embedded images are extracted directly; vector diagrams are preserved through page-level figure renders when captions are detected.
• OCR status: false
• Downstream PDF analysis status: false

Gaps

• No diagram image gaps detected.