The illustration was generated using Midjourney and upscaled and enhanced by using Magnific AI, which is an AI upscale and enhancer tool that increases and reimagines details based on prompts and parameters. It invents additional details in the image while simultaneously doubling the image resolution at each step.
View the illustration in full screen by clicking on the images.
A comparison of the generated original and the finished illustration is at the bottom of the page.
The first upscaling doubles the image resolution and significantly enhances the details.
The second upscaling allows zooming into the image while maintaining the same sharp image quality as the previous picture. This process is repeated three more times to create a close-up of the face.
At each step, additional details were reinvented.
For the neural filters to generate the desired content, some retouching was necessary for optimal quality. For example, the hands were originally generated with such low quality that it was necessary to retouch them and make a better starting point before the image was upscaled and refined with Magnific AI for the first time.
At last a final 2x upscaling and enhancing in Magnific AI for the closeup of the face.
The image resolution has now increased 11 times. The generated image from Midjourney was 1456 x 816 pixels and the finished illustration is equivalent to 16384 x 9182 pixels when zoomed in on the face.
The image from Midjourney was initially upscaled in Photoshop to 2048 x 1148 for the initial working sketch and then doubled four times with Magnific AI.
Composition showing the original generated image from Midjourney alongside details of the finished result.