Tools

Generating Images with Artificial Intelligence: Midjourney Tutorial

Learn how to create stunning images with AI to transform the communication of your liberal practice ✅ DALL-E Tutorial | Midjourney | Lexica

Working with Ourama
Generating Images with Artificial Intelligence: Midjourney Tutorial

The speed of development of artificial intelligence tools is impressive. You can quickly feel overwhelmed, so here is a tutorial to get started with image generation tools like DALL-E and Midjourney. These tools for generate high quality images Are not only Accessible, but can also respond to numerous professional issues.

Why learn to generate images for a liberal firm?

Communicate about your business

Use AI-generated images Allows you to create custom visuals For your communication campaigns. Whether for Newsletters, conferences, webinars, blog posts Or publications on your social networks, of quality images can greatly improve theinvolvement And theimpact of your messages.

Illustrate your website

Les pictures are a key part of everything website. By generating your own visuals, you can insure a graphic coherence And a professional aesthetics, thus strengthening theimage of your office.

Prompt engineering: Key to image generation

What is a prompt?

One Prompt Is a text instruction that you give to an AI for generate an image. La quality And the precision of this instruction largely determine the quality of the image generated.

Lexica: An AI-generated image bank

Lexica is a platform where you can find AI-generated images, accompanied by Prompts used to create them. This allows you to:

  • Find directly the images you are looking for.
  • You inspire existing prompts for your own creations.

Prompt engineering best practices

  1. Be specific : The more your description is detailed, the more the image generated will be Close to what you want.
  2. Use specific adjectives : Describe the colours, the textures, the shapes, and the styles.
  3. Experiment : Do not hesitate to adjust and test different prompts to get the desired result.

Prompt example : “Gray futuristic vehicle in focus, person standing diagonally behind the car, industrial look, ice world, dystopia, cold, gloomy atmosphere, people wearing storm masks, high-end workwear clothing in black, high-end workwear clothing in black, balenciaga, futuristic, realistic shooting”

Using DALL-E

DALL-E is integrated directly into GPT chat (with paid subscription), allowing you to generate images quickly and easily.

Benefits of DALL-E

  • Accessibility : Can be used directly via ChatGPT, DALL-E becomes a daily visual assistant.
  • Versatility : In addition to image generation, you benefit from all the ChatGPT features.

Disadvantages of DALL-E

  • Variable quality : Despite well-built prompts, the generated images can sometimes lack quality.

Using Midjourney

Midjourney is renowned for the exceptional quality of his images. Although thesubscription Either to 20€ per month, it offers a interesting flexibility and the possibility of building images of rare quality.

Midjourney benefits

  • High quality images : The visuals generated are often extremely detailed and aesthetics.
  • Subscription flexibility : You can subscribe for one month, generate a large quantity of images, and then cancel the subscription.

Disadvantages of Midjourney

  • Initial handling : The use of Discord Maybe baffling for new users, but a more intuitive interface is under development.

How does Midjourney work

Midjourney is the most effective platform for generating high quality images from text descriptions.

Step 1: Write a prompt

As said before, a Prompt is a detailed text description of the image you want to generate. For example, if you want an image of a consultant in a modern office, your prompt could be: “A young consultant with black hair and a short beard sits in his office at his desk, photo taken by Leica M6, realism”. The more accurate your prompt is, the closer the result will be to what you expect.

Step 2: Image Generation

After submitting your prompt, Midjourney will generate four images in less than a minute. These images are based on the details you provided and offer different variations of your original description.

Step 3: Refine the images

Once the initial four images are generated, you have several options for refining and improving these visuals:

  • Resolution improvement : You can choose one of the images and ask Midjourney to improve it, increasing its resolution and quality.
  • Zoom out : This feature allows you to see a larger version of the image, adding more context or detail around the main subject.
  • Partial change : You can select a specific part of the image for further editing or detailing.

Step 4: Use reference images

Midjourney also allows you to start from a reference image. For example, if you have a photo of your team or offices, you can use that as a starting point. Midjourney will generate images based on this reference, while also incorporating the details from your prompt. This is particularly useful for creating visuals that match your visual identity or when you like a composition.

You cannot natively generate images of yourself, for that you must use a tool called Inswapper. A bit more technical to install, I'll leave you A doc here!

Subscription flexibility

The Midjourney subscription is 20€ per month, but it offers interesting flexibility. You can:

  • Subscribe for one month, generate a large quantity of images, then cancel the subscription.
  • Come back anytime to create new visuals when you need them.

Interface and handling

Initially, Midjourney uses Discord for its user interface. Although this can be confusing for new users, the platform is working on a dedicated interface that is more intuitive and user-friendly. On Discord, you:

  • Join the Midjourney server.
  • Enter your prompt in the dedicated channel.
  • Receive and interact with your images directly in Discord.

To better understand how this works, watch our video tutorial on the subject above!