Copy link to clipboard
Copied
I asked this question wrong in the past and got a good answer but it was to the wrong question.
I have an image that displays text. But is there a way to save the image in a format that makes it impossible for computers to know that the text is there? In other words, I want people to know the text is there and be able to read it by looking at the image, but I don't want a computer to be able to identify the text in the image.
Purpose - Facebook sometimes stops ads that have 'too much text' from going forward. Meanwhile, some of the most effective Facebook ads are images with text in them. I've brought this to Facebook's attention, but they don't seem to care, I even forwarded them all the ads that are almost all text. So I have to assume that these images are being missed by Facebook because they're being saved in a format that the facebook bots can't recognize.
No! If a human can read text in the image I'm sure some image AI software could detect that text in the image how would you prevent Facebook from using AI software like that. Facebook can change the software they use anytime. If you fool them today who knows if you'll be able to fool the tomorrow. If Facebook does not care today why would the care tomorrow. If it violates their terms of use like none of this or that or advertisement whatever terms Facebook may have. The Images with te
...Copy link to clipboard
Copied
No! If a human can read text in the image I'm sure some image AI software could detect that text in the image how would you prevent Facebook from using AI software like that. Facebook can change the software they use anytime. If you fool them today who knows if you'll be able to fool the tomorrow. If Facebook does not care today why would the care tomorrow. If it violates their terms of use like none of this or that or advertisement whatever terms Facebook may have. The Images with text are not Facebook properties they are the creators property. However, it is Facebook web site and they have the right to make and have terms of use. If you violated their terms your should not be surprised if your stuff gets kicked off.
Copy link to clipboard
Copied
JJMack is correct. There are no secret hidden pixels or an actual text string in your image files that would tell FB that there is text. And it's nothing new. Commercial OCR has been around for over 20 years now and today is bolstered by additional AI algorithms. No way to cheat here. From personal experience I would say, though, that it's not actually the amount of text but rather how many pixels it occupies compared to the rest of the content. This can often be remedied by simply using different fonts or even trivial stuff like using gradated colors in the type and not large solid areas of the same color. also make sure that the main headline text/ logo text is not outweighed by additional info text. This is often interporeted as attempted manipulation of the search engines when they recognize the actual text and store it in their databases. No different than extraneous invisible HTML text being mostly prohibited these days.
Mylenium
Find more inspiration, events, and resources on the new Adobe Community
Explore Now