State of the art-ful AI: Happy Christmas!

From AI to AI and back again

This is the final entry in our ‘State of the art-ful AI’ series, where we upload a photo to Chat GPT 4 and ask it for a photo description, then ask DALL-E 3 to generate an image from the description. How similar will they be?

May your holidays sparkle with joy and laughter. May the festive glow of the season shine in your heart as it does in the lively streets. May the serene wisdom and tranquility of winter’s nature inspire peace and happiness in your home. Wishing you a magical Christmas filled with wonder and delight, just like the enchanting lights and the timeless serenity found in the simple beauty around us.

Merry Christmas! ✨

With love from and with thanks to ChatGPT 4 in December 2023 🙂

Why this workflow?

Less than a year after its launch, ChatGPT and DALL-E have already seen multiple iterations. In October 2023, ChatGPT 4 gained the capability to describe uploaded photos. To track the progress of generative AI, we uploaded these AI descriptions to the Dall-E 3 image generation model to see how faithful its output is to the original photo. A variety of photos have been selected which we’ll bring you throughout November and December.

Our original photo

This festive image was taken in Covent Garden market, London, England around 2015.

The photo as described by Chat GPT 4 in 150 words

This image shows a bustling indoor market with a glass ceiling through which natural light is streaming in, highlighting the festive decorations. Large, glossy red and purple ornaments, as well as silver disco balls, are suspended from the ceiling, creating a cheerful, holiday-like atmosphere. The architecture suggests a historic building, with the arching ironwork of the ceiling and the classic brick facade of the shops adding to the traditional charm. Below the decorations, people can be seen milling about, with some seated at a cafe on the right.

The shops, featuring signs like “CARAT” and “Venchi”, imply a variety of offerings, from jewelry to gourmet treats. The bottom of the image shows a balcony decorated with garlands, hinting that this place is likely decked out for a holiday season, possibly Christmas, considering the color scheme and the presence of a small kiosk labeled “soups & stews”, suggesting a wintry menu.

Photos generated by Dall-E 3 using the AI description

Spot the difference

Using a separate chat, we asked Chat GPT what differences it finds between the original photo and the FIRST image generated above:

  1. Lighting: The first image has a more natural and softer lighting, while the second image appears to have a warmer, more artificial lighting that highlights the colors and details of the scene.
  2. Color Saturation: The colors in the second image are more vibrant and saturated compared to the first image where the colors are more muted.
  3. Sharpness: The second image is sharper, with more defined details, whereas the first image has a softer focus.
  4. Perspective and Composition: The first image is taken from a lower angle and closer to the baubles, making them appear larger. The second image is taken from a higher angle, giving a more expansive view of the location.
  5. Decoration Density: There seem to be more decorations, especially garlands, visible on the balconies in the second image compared to the first one.
  6. Shop Names: The shop names and details are more legible in the second image due to the increased sharpness and lighting.
  7. People: There are more people visible in the second image, and they are more in focus, adding to the liveliness of the scene.
  8. Structural Elements: Architectural details, like the ironwork of the balconies and the glass ceiling, are more pronounced in the second image.
  9. Color Palette: The overall color palette of the second image leans towards warmer tones, while the first image has a cooler, more neutral palette.
  10. Editing: The second image appears to have undergone more post-processing, potentially including HDR (High Dynamic Range) techniques, to enhance contrast and detail.

Example workflow video

Here’s our workflow in action, using an ice cream truck image. This is not an explainer video – It shows the main steps of the above ChatGPT 4 workflow in November 2023.