APACE-EU

Ask the Expert: Images and Accessibility

The Q&A of the Ask the Accessibility Expert session with Kirsi Ylänne about images and accessibility.
20 May 2025


In this session host Katie Durand and accessibility expert Kirsi Ylänne from the Accessibility Library Celia in Finland discuss the challenges of complex images and share practical advice and actionable best practices for creating accessible images and image descriptions. 

The expert

Kirsi Ylänne is an accessibility specialist at the Accessibility Library Celia. In the past she worked as a production manager from 2005 to 2014. She is currently a standing committee member of the IFLA (International Federation of Library Associations) in the Libraries serving persons with print disability section. She served on that from 2015 to 2023, and she was the section chair from 2017 to 2021. She was also a member of the Finnish Braille Authority from 2006 to 2018.

Questions and answers

How important is it to mention and describe the colours that are used in this graph? And are there some cases where all this use of colour might cause sensory overload?

With for instance a bar diagram or some charts you could refrain from mentioning the colours unless if there's a reference in the body text to the colour. If in the body text you can see blue lines, then it would be important to mention which is the blue line, but otherwise concentrate on the on the diagrams and charts on what the colour represents. Don’t go into every detail describing the colours so the image description can be shorter and there isn’t an information.

What is the difference between alt text and a long image description?

When we talk about alt text, we mean short descriptions of images. An alt text is only accessible for screen reader users. There is no visual text. It's hidden in the code of the ebook, and the length is usually like one to 3 sentences. An alt text cannot be formatted, it's just text. 

With long descriptions or extended description as it's sometimes called, it's a text, and it is available for all readers. The text in the long description can be formatted. So you can use headings. You can use emphasis, list, or tables. There are different techniques to add long descriptions. One way is to add a link to a separate html file where the description is located, and there has to be also the link back to the body text. If you don't have that backlink, then then you end up having just a separate html file. Another possibility is to add a description using a details element. This is a handy way for publications that are published as a website. But for epubs, it's not so good because many epub readers paginate the content. And if you have this extended details element that can be extended, then the pagination is not working very well. So the safest way is to use the separate html file  

With pdfs it's more complicated. For long descriptions in pdf you need have an appendix where you collect all the long descriptions and then use links to the long description and back to the body text that help the screen reader user to navigate inside the pdf. But it really depends on the purpose of the image what kind of descriptions are needed. Sometimes a short alt text is enough, but sometimes a longer description is needed too. 

How do you actually go about building a long description?

You have to remember that when you write a long description, you don't have to write every little detail of the image. The aim is to give the relevant information of the image for the reader who may not see or understand the image and remember that the long description is available for anybody.  

The context is always the most important aspect. What's the purpose of the image? And what is the target audience? What kind of language can you use for the target audience? And what is the suitable tone and language, and of course, use the same kind of language and tone that is in the body text. If it's scholarly literature, you can use terms that you can assume the target audience knows. But if you are doing a book for for 1st graders, you cannot have any new terms in the long description, and it's good to start from the generic. So you have the kind of overview of the image and then go to the details. 

And since you can use formatting in long description, you can give a good structure and use headings and, if necessary, subheadings. Also use list and tables, but don't overdo it. And of course always make sure that there is the return link to the body text. 

What style of language can you use? Is it possible to use metaphors, or maybe some sort of creative language in certain contexts?

If it’s the tone that's used in the body text, and you can assume that the target audience understands that.  

What is the opinion on image descriptions generated by artificial intelligence? And are there any tools that are recommended? 

When we assume that the copyright issues are okay, that you are using artificial intelligence in a safe way, in a safe environment and not violating any copyright issues, AI can certainly be helpful. Artificial intelligence is doing quite well with charts and diagrams already. So you can save a lot of time using it, but don't trust that it does the job for you totally. You always have to check what the result is. And especially if you are not doing it in English.  

If you create the long description for a digital book and make it visible to everyone, do you have to do the same for the print version?

The European Accessibility Act only concerns ebooks. It is up to the publishers. But you might have a kind of a quality check on the manuscript and the kind of text, and maybe think about it when you re-read an old image description. Is there something that you could add to the body text? Is it something that is so important that should be already in the body text.

What guidance is there for writing table summaries?

If you make accessible tables, you don't need the table summaries. If you have a really complex table, then a summary that would explain the structure of the table is recommended. But try to make the tables as simple as possible. Could a table be divided into several tables? And, if possible, try with the screenreader, so you can navigate the table and understand it.  

Can you use just alt text or do you have to use extended description when the original image is too complicated to be made accessible.

If you cannot improve the image, then provide an atl text and a good quality long description. That would be enough then.

see all


03 June 2025

Ask the Expert: EPUB fixed layout

The Q&A of the Ask the Accessibility Expert session with Gregorio Pellegrino about fixed layout EPUBs and accessibility.

20 May 2025

Ask the Expert: Images and Accessibility

The Q&A of the Ask the Accessibility Expert session with Kirsi Ylänne about images and accessibility.

25 February 2025

Ask the Expert: Reading apps and accessibility

A summary of the Ask the Accessibility Expert, including the Q & A with Richard Orme about the accessibility features of reading apps.