Adobe Firefly, Adobe Photoshop
Adobe Firefly is Adobe’s generative technology. It uses the Adobe Firefly vector model, which can directly generate vector images (rather than raster image formats like JPEG or PNG). Vector images can then be easily processed in tools like Adobe Photoshop etc.
Firefly has been trained with public domain images and as such has kept any copyright issues away. It is a diffusion model and Adobe keeps technical details somewhat unexplained, but the main advantage is that it can generate vectors which can be edited with Adobe tools for fine details.
I’ve used Firefly through Adobe Photoshop and generated some images through text prompts. The generation is done by selecting an area on the canvas and using a textual description to fill the selected area with a generated image.
Playground.ai
This is another OpenAI tool that is alternatively named gpt-3.5-turbo. It gives you 500 images per day with some limitations in generating (about the size of the image, lack of multiple images per prompt etc.). I gave it a prompt and it generated the image but without using all the details in my description, although the generated image is not bad. Note that DALL-E 3 uses ChatGPT as its front-end and as such the prompts internally generated to be the input for image generation are much better for DALL-E.
Mage.space
This is a site that has basic models but can give you access to 160+ models if you go for Pro membership. The basic models available are SDXL, Stable Diffusion 1.5 and Stable Diffusion.
I provided the following prompt and got a reasonable image.
create an image in ottoman miniature style, depicting two foreigners in istanbul wearing European clothes and trying to meet covertly to execute a secret plan
Craiyon
This was formerly called DALL-E Mini. I gave it a prompt to create an oil painting in the style of 16th-century Dutch painters with the following prompt and got the resulting 256x256 image. It wasn’t possible to enlarge it to 1024x1024 or another bigger size.
You are a painter in the Netherlands in the 17th century. Create an oil painting that depicts three small girls running around a windmill, with a dog accompanying them
I gave the same prompt to DALL-E 3 to compare the results. The resulting 1024x1024 picture looks much better.
YouImagine
This is an application that uses the Stable Diffusion XL and XL HD models for image generation. It has a free mode and a couple of styles to define. It gives free mode users 10 images to generate, to try out the system.
monk with two katana swords getting ready to fight two robots with a knife in each hand, spectators watching the fight with horror, flying vehicles in the background
Partly successful. You can also combine YouChat with YouImagine to give a chat/language front-end to the tool. Here is what YouChat produced for the prompt given above for YouImagine.
In a dramatic scene, a monk stands tall, holding two katana swords with a focused expression on his face. He is dressed in traditional monk robes, which flow around him as if caught in a gentle breeze. The monk's muscles are tense, ready for action.
Facing the monk are two formidable robots, each wielding a knife in each hand. The robots are sleek and metallic, with sharp edges and glowing red eyes. They stand in a menacing posture, ready to engage in combat.
Surrounding the scene are horrified spectators, their faces filled with shock and fear. They watch the impending battle unfold with bated breath, unable to tear their eyes away from the intense confrontation.
In the background, futuristic flying vehicles soar through the sky, adding to the sense of a dystopian world where technology and ancient martial arts collide.CandyIcons
This is actually a web site that can generate iOS or web icons from a given image, but it also has an AI mode (with the paid option) that can be used to produce these icons automatically.
Looka
This is a Logo design application that starts with some keywords from the user and the selection of a template, then proceeds with the production of a Logo kit, which has lots of promotional materials designed. The kit can be purchased and downloaded. Although it requires a subscription, saved graphical assets will be owned by the user afterwards.
(to be continued…)