Re: PDF - Text Extraction File

On 20/10/2025 23:07, Adam Cooper wrote:
> applications out there that can vectorise a raster image

The document appears to have been authored in vector form. As already 
noted, being in vector form, even with text as text, is not enough.

There are no raster images contributing to the final visual result, 
although there there is content that has been "painted" over, which may 
explain the presence of SVG mask elements and possibly other elements, 
with bitmapped images.

They've added this (slightly redacted here) as ordinary text, at the 
very top:

"Accessibility note: This is a detailed map of Ivany Campus. If you need 
support navigating the campus, please call Student Services at 902-491-...."

According to Google's AI, neither Adobe Illustrator, nor Inkscape (which 
I used for the forensics) can create tagged PDF, which would be needed 
for a PDF version to have alternative text.  However, the SVG from 
Inkscape displays visually correctly in Firefox, and it would be 
possible to, at least, hand edit in alternative text in that SVG.  (I 
assume that there is currently no tagging in the PDF, given that Adobe 
Illustrator doesn't support this, but I don't have an open source tool 
that I know to support tagged PDF.)

LibreOffice does have some tagged PDF support, but it cannot read the 
PDF completely accurately (although it is happy with the SVG created by 
Inkscape), and I think tagged PDF support is limited to the Writer mode, 
whereas PDFs import into the Draw mode.

Received on Tuesday, 21 October 2025 17:55:45 UTC