Keep the original PDF intact
Retain the original file and record its identity before producing page images or crops. Keep the page order, page count and any available document metadata with the source record. A derived image is a working representation of a page; it should remain distinguishable from the original PDF.
Inspect the type of page content
Check whether the PDF contains scanned pages, embedded raster marks or native document content. A visible signature can be a handwritten mark captured in a scan or an image inserted into a document. Appearance alone does not establish how or when that mark was placed there.
Check the page image quality
Use the clearest available source. Avoid repeated screenshots and unnecessary recompression. Check that the whole signature and useful surrounding context are present, including nearby text, lines or stamps that may affect extraction. If an image export is necessary, record the page number, resolution and export settings.
Review candidate regions
The document extraction workflow accepts PDF pages and supported image inputs. It locates candidate signature-shaped regions and produces individual crops for further work. A candidate can include nearby print, ruling or a stamp; check the crop boundaries and keep the source page available when adjusting a working crop.
Keep specimens connected to their pages
Maintain a separate record for each extracted region. The page number, crop location, file identity and relationship to the source document make the specimen usable in later comparison and reporting. A clean crop should not replace the page-level context or erase the distinction between multiple marks on one page.
Choose the next analysis by its accepted input
Some comparison workflows accept PDFs directly, while others require signature images. Check the selected workflow’s input requirements before transferring a crop. Prepare questioned and known material consistently, retain the sealed source and identify each conversion in the processing record.