With AI, you will stand a much better chance if you break down your image into smaller pieces.
That helps in many ways:
- AI will probably not generate a full size book cover at 300dpi, which is the recommended minimum for printing
- You stand a better chance of getting good faces if there are only a few people in each image
- Likewise, AI is likely to do something closer to what you ask if you don’t ask it to do too much in each image.
- You can work individually on each of the images until you get the best result, without upsetting other parts of the image
- AI cannot usually generate sensible titles, but a book without a title isn’t going to fly off the shelves
- Once you have all the elements, a conventional graphics package can take over to cut out the elements, enhance the resolution, adjust the relative sizes and arrange them in a pleasing layout leaving space for the titles.
For example:
