๐๏ธ ๐ข Introduction
Figuring out the best prompt to create a perfect image is a particular challenge. Research into methods to do so is not quite as developed as text prompting. This may be due to inherent challenges in creating objects which are fundamentally subjective and often lack good accuracy metrics. However, fear not, as the image prompting community(@parsons2022dalleprompt) has made great discoveries about how to prompt various image models(@rombach2021highresolution)(@ramesh2022hierarchical).
๐๏ธ ๐ข Style Modifiers
Style modifiers are simply descriptors which consistently
๐๏ธ ๐ข Quality Boosters
Quality boosters(@oppenlaender2022taxonomy) are terms added to a prompt to improve
๐๏ธ ๐ข Repetition
Repeating the same word within a prompt, or similar phrases can cause the model
๐๏ธ ๐ข Weighted Terms
Some models (Stable Diffusion, Midjourney, etc.) allow you to weight terms in a prompt. This can be used to emphasize certain words or phrases in the generated image. It can
๐๏ธ ๐ข Fix Deformed Generations
Deformed generations, particularly on human body parts (e.g. hands, feet), are a common issue with many models. This can be dealt with to some extent with good negative prompts(@blake2022with). The following example is adapted from this Reddit post.
๐๏ธ ๐ข Shot type
Shot types refer to the various camera angles and compositions used in filmmaking to capture different perspectives and visual effects within a scene. These shot types play a crucial role in conveying the story, setting the mood, and influencing the audience's perception and emotions.
๐๏ธ ๐ข Midjourney
Midjourney is an AI image generator that operates through a Discord bot interface as well as a web app (an API version of Midjourney is planned). The process for generating images with Midjourney follows the same fundamental principles as other AI image generators, including the use of prompts to guide the generation process.
๐๏ธ ๐ข Resources
The image prompt engineering field is very diverse. Here is a good list of