Simplify adding Alternative Text for images SOLVED

[Jensen Head]

Ideally, I would like to see the use of a cloud-based tool for automatically analyzing image content using a neural network and filling in non-empty values ​​of this attribute for images of a user-specified minimum size in pixels.

An example of automatically generated alternative text — docs.google.com/document/d/1JRlpx9jPr652IwBNK1lzMATm8YnBnbk71FKOxPzVAOE/edit

However, until the app gets online features, it would be nice to at least see an "Alternate text" field in the image properties list. Currently, it only shows the following attributes:

§ General
Layer
§ Style
Opacity
Blend Mode
§ Image Details
Width
Height
Type
BPC
Compression
Color Space
Mask Type
Object Number
Layer

I did not find a corresponding item in the context menu of images on the page image and in the Thumbnails Pane.I didn't find any hints on how to do this in PDF-XChange Editor V10 User Manual (13/03/2025). If I understand correctly, the "Set Alternate Text" command (cmd.acc.alt.text) does something different.

You can always verify the need for alternative text by trying to read a PDF document without descriptions of illustrations, but filled with formulas, photographs or graphs, with your eyes closed, using only tools to assist the visually impaired. This is an important accessibility tool. As a bonus, you will get better indexing of the document in search engines.

[Daniel - PDF-XChange]

Hello, Jensen Head

i think you have come to the realization yourself in another discussion, but just incase, the accessibility data is not a part of the image itself and so cannot be part of the image properties.

With that said, we do offer a simplified way to set alt text for images, the button is located on the Accessibility tab:

image.png

It presents a dialogue for you to select each image that is present, and supply the alt text for it directly.

Kind regards,

[Jensen Head]

I click the Set Alternate Text (cmd.acc.alt.text) button in Select Text (cmd.tool.selectText) and Images (cmd.tool.editContent.images) modes with an image selected, but nothing happens. What am I doing wrong? The document is not digitally signed, is not in PDF/A, PDF/X, PDF/E, PDF/UA, or PDF/VT formats, and is not password protected.

PDF-XChange Editor 10.6.0, build 396

[Daniel - PDF-XChange]

Hello, Jensen Head

Could I ask for a copy of a PDF you have where this does not work? Any figures I have tagged seem to properly prompt me to adjust the alt text when I use this button.

image.png

Kind regards,

[Jensen Head]

I tried selecting the image in the Content Pane first, but there is still no response to pressing the button you specified.

[Daniel - PDF-XChange]

Hello, Jensen Head

Thank you, I see the issue here. You havent actually specified whether this content is intended to be considered an "image", "figure", "background/artifact", or some other type of content (it could even be a photograph of a text block, in which case, you could possibly give it the paragraph tag).
To do so, use the "reading order" tool, and drag a box around the image (my screenshot's zoom level made it awkward to do so, I advise zooming out more, so you can easily get outside the image bounds):

image.png

Then select the option in question. In this case, "image"

image(1).png

Or if you wanted to, flag it as a paragraph, because, I feel like pretending this is an image of a block of text now!

image(2).png

As a slight afterthought... I am honestly not 100% certain if my final example there (with the paragraph tag) is an "appropriate" use of the tagging feature. Since I do not need to submit documents I use this feature with anywhere... I haven't personally encountered any reports of it being "improper" either.
In most cases I would recommend that you stick to labelling images as "figure" or "background object" predominantly, then using either the properties pane (with the selection from the tags window), or the "set alt text" button, to properly format those items. But since this is "technically possible" and it prevents the "your figures have not been labelled" error in our own reporting, I wanted to mention it here.

Kind regards,