image description
An infographic titled “How To Write Alt Text” featuring a photo of a capybara. Parts of alt text are divided by color, including “identify who”, “expression”, “description”, “colour”, and “interesting features”. The finished description reads “A capybara looking relaxed in a hot spa. Yellow yuzu fruits are floating in the water, and one is balanced on the top of the capybara’s head.”
via https://www.perkins.org/resource/how-write-alt-text-and-image-descriptions-visually-impaired/
You must log in or register to comment.
A capybara in the library with a candlestick.
He was definitely the murderer
1 + 1 + 2 + 1
A true classic
Bro I fucking love capybaras so much
10/10 animal, fucking brilliant.
my favourite animal
the carbonara
Potentially also useful for creating good prompts for AI image generators?
It’s essentially by-hand CLIP, that’s how the training data for CLIP came into being, it was descriptive text for images.
Explains why it sucks so much shit.
CLIP is pretty decent for what it does though
It’s only useful if the AI was trained on similar prompts. A lot of the anime style ones work best with lists of tags, while the realistic ones work best with descriptions like above.
Prompts are just the reverse of image recognition AI tagging stuff.
Alt text is exactly the kind of tedious work that AI would be good at doing, but everyone in the fediverse seems to have a huge hate boner for ANYTHING AI…
Fediverse: write a fucking essay every time you post an image… But make sure you waste time doing it manually, instead of using AI tools!!!
If you have really detailed image tags, a model trained on them can make great outputs.
We don’t do that here.
Me writing alt text: Time is a flat circle. God is a sock.
Just as important is “decorative images” where you explicitly leave the alt empty https://www.w3.org/WAI/tutorials/images/decorative/
You know those little [] that appear when you upload an image? You can put alt-text in there.
Ignorant question: isn’t alt text primarily for visually impaired people? If so, what is the point of including info about color?
Color can provide useful context. For example, in the case of this image, imagine if in a thread about it there was some discussion of the ripeness of the yuzu fruit.
You can also become visually impaired at points other than birth in life, and know colours and stuff
That’s a very good point!
Is this not the kind of thing machine vision/language models would be really good at?
