Returning to Misspellings in AI-generated Photos – Bredemarket
There’s a LinkedIn consumer who writes and titles posts on numerous technical subjects…after which ruins the posts by illustrating them with an AI picture generator and together with the submit title within the picture.
Why are they ruined? As a result of some I mage mills can’t spell. For instance, one of many pictures contains the so-called phrase CYBERCEERITYY.
In order for you your AI-generated pictures to incorporate correctly-spelled phrases, both use quick easy phrases or use massive multimodal mannequin (LMM) image-text mills. See my earlier submit, “Textual content Technology in Photos? Use an LMM” on the https://bredemarket.com/2024/07/30/text-generation-in-images-use-an-lmm/ URL.
Apparently Google Gemini’s Imagen 4 incorporates LMM capabilities; I fed it a fancy textual title and it was spelled appropriately within the ensuing picture.
(Imagen 4)
Printed by
