**Created**:: [[2023-04-07 Friday]] **Link**:: https://youtu.be/Pnrizu4LKO8 **Duration**:: 6:31  # Summary In this video, the creator discusses Midjourney's innovative image description tool, which provides four distinct descriptions for uploaded images, helping users to understand the underlying prompt logic and create new artistic prompts. They demonstrate a creative experimentation process with the famous painting, Mona Lisa, leading to diverse and imaginative derivative images. Additionally, the video explores three methods combining Midjourney with external tools like Excalidraw and ChatGPT to refine and enhance image generation. # Key Takeaways - Midjourney's image description tool offers four varying descriptions that help users grasp its prompt logic better. - Creative looping with images can lead to unexpected and imaginative transformations of original artworks. - Combining images generated in Midjourney with external resources and editing tools can enhance the creative process and yield desired results. # Content Hey everyone! In this video, I'll be discussing Midjourney's latest feature, the image description tool, that returns four descriptions for any given image. This feature is super powerful as it allows you to better understand Midjourney's prompt logic and craft new image prompts using the descriptions provided by the tool. In the first part of the video, I'll explore how far I can get from a well-known painting, the Mona Lisa, by creating a feedback loop. At the end of this video, I'll share three additional ideas to help you generate creative images using image generation, image editing, and ChatGPT. So make sure you watch the entire video to not miss out on any of these insights. [* ](https://youtu.be/Pnrizu4LKO8?t=0) As I discussed in my book on a page summary for the extended mind, we are creatures that rely heavily on looping our ideas through external mediums and ourselves to solve problems and create innovative solutions. I wanted to try the creative effects of looping with Midjourney. When you submit an image to Midjourney using the describe function, it returns four different descriptions. Based on my experience, the first description is usually more precise, while the last one tends to be more creative. Although you could argue that the Mona Lisa was an easy target for Midjourney, I was pleasantly surprised when it generated much longer and more intricate descriptions instead of just saying "Mona Lisa by Leonardo da Vinci." [* ](https://youtu.be/Pnrizu4LKO8?t=55) For my first experiment, I followed a simple process where I always generated the first recommended prompt, then selected the first image out of the four, asked for a description, and generated it again. After some minor divergence, Midjourney seemed to circle back to the Mona Lisa repeatedly. So after seven generations, I stopped. Next, I took a more directed approach by selecting descriptions that felt the most divergent, and images that were the most different. This led me to some gruesome and gritty horror images, which slowly transformed into a golden mask with red paint. [* ](https://youtu.be/Pnrizu4LKO8?t=118) In my second attempt, I got lucky when Midjourney generated stained glass images, then swirls of colors, which eventually transformed into a psychedelic illustration of whimsical beauty. In my third attempt, I put lucky again when the Mona Lisa transformed into a male character, and the red robe turned white. The white hat became a Son God’s crown, which morphed into an antler crown, and finally into a wax sculpture of a king. Although I was never able to find a way to leave the portraits behind and convert the Mona Lisa into something wholly abstract or different, I still managed to get pretty far from the original image. [* ](https://youtu.be/Pnrizu4LKO8?t=200) Now let's move on to the additional ideas I promised earlier. I recently wanted an image of a bear in a nightcap waking up from its winter sleep; however, I wasn't able to convince Midjourney to generate the image I had in mind using my usual prompt tricks. So, I tried a new approach by using the describe feature in Midjourney and some external images from Flat Icon. I selected the image I liked the most from those I generated with my crafted prompts, then looked for a nightcap on Flat Icon and used the color picker to align the colors of the nightcap with the picture. Next, I used Excalidraw and Obsidian to place the cap on the bear, resulting in a pretty good image already. [* ](https://youtu.be/Pnrizu4LKO8?t=244) I then fed the image back to Midjourney and asked it to describe the image. I used these descriptions to generate images, and these resulted in some nice pictures, but most of them still didn't have the nightcap on. To try to solve this, I first asked ChatGPT to combine the four image prompts into a single prompt by asking it to distill these four image prompts and come up with a single cohesive prompt. Then, I generated an image using the combined prompt. Finally, I used my edited image—the one with the nightcap from Flat Icon—as an image prompt and combined it with the distilled prompt generated by ChatGPT. This resulted in some nice pictures, including one with the bear in bed with the nightcap on. This is what I wanted! [* ](https://youtu.be/Pnrizu4LKO8?t=346) So that's all I wanted to share with you today. I hope this gave you some inspiration to go and generate some images of your own. Please share your creative image generation workflows and the links to the pictures you've created in the comments below. Thank you! [* ](https://youtu.be/Pnrizu4LKO8?t=374) # YouTube Details ## YouTube Description In this video, I explore Midjourney's latest feature - the image description tool that returns four different descriptions of any given image. I share my journey of looping with Midjourney, starting with the well-known painting, Mona Lisa, and eventually creating some truly unique and creative images using image generation, image editing, and ChatGPT. I also share some additional ideas on how to generate creative images using Midjourney and other tools. Watch this video to get inspired and learn how to generate your own creative images. Don't forget to share your workflows and links to your creations in the comments below! -------- [00:00](https://youtu.be/Pnrizu4LKO8?t=0) Intro [00:55](https://youtu.be/Pnrizu4LKO8?t=55) Why looping is important [01:17](https://youtu.be/Pnrizu4LKO8?t=77) Midjourney's "/describe" function [01:54](https://youtu.be/Pnrizu4LKO8?t=114) The Mona Lisa Journey No.1 [02:30](https://youtu.be/Pnrizu4LKO8?t=150) The Mona Lisa Journey No.2 [02:57](https://youtu.be/Pnrizu4LKO8?t=177) The Mona Lisa Journey No.3 [03:18](https://youtu.be/Pnrizu4LKO8?t=198) The Mona Lisa Journey No.4 [03:44](https://youtu.be/Pnrizu4LKO8?t=224) Summary: The Mona Lisa Journey [04:01](https://youtu.be/Pnrizu4LKO8?t=241) Bear in a Nightcap [06:12](https://youtu.be/Pnrizu4LKO8?t=372) Please Share Your Worlflows and Images in the Comments -------- 🙏 If you find my videos helpful, please say thanks by buying me a coffee: https://ko-fi.com/zsolt 📩 If you want to connect, you can reach me: (@zsviczian) on the Obsidian Members Group (OMG) on Discord, or on 🐦 Twitter: https://twitter.com/zsviczian This video was created with the Obsidian-Excalidraw-Plugin. 🍿Watch next: -------- 📽️Design unique icons with AI: https://youtu.be/ZeqsDWBawdo 📽️Excalidraw Slideshow script: https://youtu.be/HhRHFhWkmCk 📽️The Extended Mind Book-on-a-Page: https://youtu.be/oXMKF8zqG6E 📽️Guide to Creating an Illustrated Book-on-a-Page Summary Without Reading the Book: 📽️https://youtu.be/snfn1ylocvY 🔗Links -------- 🚀Visual Thinking Workshop: https://www.visual-thinking-workshop.com/ 🎨Excalidraw Plugin: https://github.com/zsviczian/obsidian-excalidraw-plugin 📝Obsidian: https://obsidian.md 👉Flaticon: https://www.flaticon.com/ 👉Midjourney: https://midjourney.com/ 📖My blog: https://zsolt.blog 🙏Support my work: https://ko-fi.com/zsolt ## YouTube Transcript [hey everyone in this video I'll be](https://youtu.be/Pnrizu4LKO8?t=1) [discussing mid Journey's latest feature](https://youtu.be/Pnrizu4LKO8?t=4) [the image description tool that returns](https://youtu.be/Pnrizu4LKO8?t=7) [four descriptions for any given image](https://youtu.be/Pnrizu4LKO8?t=10) [this feature is super powerful as it](https://youtu.be/Pnrizu4LKO8?t=13) [allows you to better understand my](https://youtu.be/Pnrizu4LKO8?t=16) [Journey's prompt logic and craft New](https://youtu.be/Pnrizu4LKO8?t=19) [Image prompts using the descriptions](https://youtu.be/Pnrizu4LKO8?t=21) [provided by the tool](https://youtu.be/Pnrizu4LKO8?t=24) [in the first part of the video I'll](https://youtu.be/Pnrizu4LKO8?t=26) [explore how far I can get from a](https://youtu.be/Pnrizu4LKO8?t=28) [well-known painting Mona Lisa by](https://youtu.be/Pnrizu4LKO8?t=31) [creating a feedback loop at the end of](https://youtu.be/Pnrizu4LKO8?t=33) [this video I'll share three additional](https://youtu.be/Pnrizu4LKO8?t=37) [ideas to help you generate Creative](https://youtu.be/Pnrizu4LKO8?t=39) [Images using image Generation image](https://youtu.be/Pnrizu4LKO8?t=42) [editing and chai GPT](https://youtu.be/Pnrizu4LKO8?t=45) [so make sure you watch the entire video](https://youtu.be/Pnrizu4LKO8?t=48) [to not miss out on any of these insights](https://youtu.be/Pnrizu4LKO8?t=50) [as I discussed in my book on a page](https://youtu.be/Pnrizu4LKO8?t=55) [summary for the extended mind they are](https://youtu.be/Pnrizu4LKO8?t=57) [creatures that rely heavily on looping](https://youtu.be/Pnrizu4LKO8?t=60) [our ideas through external mediums and](https://youtu.be/Pnrizu4LKO8?t=63) [ourselves to solve problems and create](https://youtu.be/Pnrizu4LKO8?t=66) [innovative solutions](https://youtu.be/Pnrizu4LKO8?t=69) [I wanted to try the creative effects of](https://youtu.be/Pnrizu4LKO8?t=72) [looping with my journey](https://youtu.be/Pnrizu4LKO8?t=74) [when you submit an image to my journey](https://youtu.be/Pnrizu4LKO8?t=78) [using the describe function it returns](https://youtu.be/Pnrizu4LKO8?t=81) [four different descriptions](https://youtu.be/Pnrizu4LKO8?t=85) [based on my experience the first](https://youtu.be/Pnrizu4LKO8?t=87) [description is usually more precise](https://youtu.be/Pnrizu4LKO8?t=90) [while the last one is more creative](https://youtu.be/Pnrizu4LKO8?t=92) [although you could argue that Mona Lisa](https://youtu.be/Pnrizu4LKO8?t=96) [was an easy target for my journey I was](https://youtu.be/Pnrizu4LKO8?t=99) [pleasantly surprised when it generated](https://youtu.be/Pnrizu4LKO8?t=102) [much longer and more intricate](https://youtu.be/Pnrizu4LKO8?t=105) [descriptions instead of just saying Mona](https://youtu.be/Pnrizu4LKO8?t=107) [Lisa by Leonardo da Vinci](https://youtu.be/Pnrizu4LKO8?t=111) [for my first experiment I followed a](https://youtu.be/Pnrizu4LKO8?t=115) [simple process where I always generated](https://youtu.be/Pnrizu4LKO8?t=118) [the first recommended prompt then](https://youtu.be/Pnrizu4LKO8?t=121) [selected the first image of the four](https://youtu.be/Pnrizu4LKO8?t=124) [asked for a description and generated it](https://youtu.be/Pnrizu4LKO8?t=128) [again](https://youtu.be/Pnrizu4LKO8?t=131) [after some minor Divergence](https://youtu.be/Pnrizu4LKO8?t=133) [mid-journey seemed to Circle back to](https://youtu.be/Pnrizu4LKO8?t=137) [Mona Lisa repeatedly](https://youtu.be/Pnrizu4LKO8?t=140) [so after seven generations](https://youtu.be/Pnrizu4LKO8?t=143) [I stopped](https://youtu.be/Pnrizu4LKO8?t=146) [next I took a more directed approach](https://youtu.be/Pnrizu4LKO8?t=150) [selecting descriptions that felt the](https://youtu.be/Pnrizu4LKO8?t=153) [most Divergent and images that were the](https://youtu.be/Pnrizu4LKO8?t=157) [most different](https://youtu.be/Pnrizu4LKO8?t=161) [this led me to some gruesome and gritty](https://youtu.be/Pnrizu4LKO8?t=163) [horror images which slowly](https://youtu.be/Pnrizu4LKO8?t=167) [transformed into a golden mask with red](https://youtu.be/Pnrizu4LKO8?t=171) [paint](https://youtu.be/Pnrizu4LKO8?t=175) [in my second attempt I got lucky that](https://youtu.be/Pnrizu4LKO8?t=177) [mid Journey generated stained glass](https://youtu.be/Pnrizu4LKO8?t=181) [images](https://youtu.be/Pnrizu4LKO8?t=184) [then swirls of colors which eventually](https://youtu.be/Pnrizu4LKO8?t=185) [transformed into a psychedelic](https://youtu.be/Pnrizu4LKO8?t=189) [illustration of a Whimsical Beauty](https://youtu.be/Pnrizu4LKO8?t=192) [in my third attempt](https://youtu.be/Pnrizu4LKO8?t=198) [I put lucky again when Mona Lisa](https://youtu.be/Pnrizu4LKO8?t=200) [transformed into a male character](https://youtu.be/Pnrizu4LKO8?t=203) [and the red robe turned white](https://youtu.be/Pnrizu4LKO8?t=206) [the white hat became a song God's Crown](https://youtu.be/Pnrizu4LKO8?t=210) [which turned into an antler crown](https://youtu.be/Pnrizu4LKO8?t=215) [and finally a vax sculpture of a king](https://youtu.be/Pnrizu4LKO8?t=219) [although I was never able to find a way](https://youtu.be/Pnrizu4LKO8?t=225) [to leave the portraits behind and](https://youtu.be/Pnrizu4LKO8?t=228) [convert Mona Lisa into something wholly](https://youtu.be/Pnrizu4LKO8?t=231) [abstract or different](https://youtu.be/Pnrizu4LKO8?t=234) [I still managed to get pretty far from](https://youtu.be/Pnrizu4LKO8?t=236) [the original image](https://youtu.be/Pnrizu4LKO8?t=239) [now let's move on to the additional](https://youtu.be/Pnrizu4LKO8?t=242) [ideas I promised earlier](https://youtu.be/Pnrizu4LKO8?t=244) [I recently wanted an image of a bear in](https://youtu.be/Pnrizu4LKO8?t=247) [a night cap waking up from its winter](https://youtu.be/Pnrizu4LKO8?t=250) [sleep](https://youtu.be/Pnrizu4LKO8?t=252) [however I wasn't able to convince my](https://youtu.be/Pnrizu4LKO8?t=253) [journey to generate the image that I had](https://youtu.be/Pnrizu4LKO8?t=256) [in mind using my usual prompt tricks](https://youtu.be/Pnrizu4LKO8?t=259) [so I tried a new approach by using the](https://youtu.be/Pnrizu4LKO8?t=262) [describe feature in mitchernly and some](https://youtu.be/Pnrizu4LKO8?t=265) [external images from Flat icon](https://youtu.be/Pnrizu4LKO8?t=268) [I selected the image I liked the most](https://youtu.be/Pnrizu4LKO8?t=271) [from those that I generated with my](https://youtu.be/Pnrizu4LKO8?t=273) [crafted prompts then I looked for a](https://youtu.be/Pnrizu4LKO8?t=276) [night cap on flat icon and used the](https://youtu.be/Pnrizu4LKO8?t=279) [Color Picker to align the colors of the](https://youtu.be/Pnrizu4LKO8?t=282) [night cap with the picture](https://youtu.be/Pnrizu4LKO8?t=285) [next I used excalib raw and obsidian to](https://youtu.be/Pnrizu4LKO8?t=288) [place the cap on the bear resulting in a](https://youtu.be/Pnrizu4LKO8?t=292) [pretty good image already](https://youtu.be/Pnrizu4LKO8?t=296) [I then fed the image back to my journey](https://youtu.be/Pnrizu4LKO8?t=298) [and asked it to describe the image](https://youtu.be/Pnrizu4LKO8?t=301) [I used these descriptions to generate](https://youtu.be/Pnrizu4LKO8?t=307) [images](https://youtu.be/Pnrizu4LKO8?t=310) [and these resulted in some nice pictures](https://youtu.be/Pnrizu4LKO8?t=313) [but most of them still didn't have the](https://youtu.be/Pnrizu4LKO8?t=315) [night cap on](https://youtu.be/Pnrizu4LKO8?t=318) [to try to solve this I first asked Chad](https://youtu.be/Pnrizu4LKO8?t=320) [GPT to combine the four image prompts](https://youtu.be/Pnrizu4LKO8?t=324) [into a single prompt by asking Che GPT](https://youtu.be/Pnrizu4LKO8?t=327) [to distill these four image prompts and](https://youtu.be/Pnrizu4LKO8?t=331) [come up with a single prompt](https://youtu.be/Pnrizu4LKO8?t=335) [then I generated an image using the](https://youtu.be/Pnrizu4LKO8?t=338) [combine prompt](https://youtu.be/Pnrizu4LKO8?t=341) [finally](https://youtu.be/Pnrizu4LKO8?t=344) [I used my edited image the one with the](https://youtu.be/Pnrizu4LKO8?t=346) [night cap from Flat icon as an image](https://youtu.be/Pnrizu4LKO8?t=349) [prompt and combined it with the](https://youtu.be/Pnrizu4LKO8?t=352) [distilled prompt generated by Chain GPT](https://youtu.be/Pnrizu4LKO8?t=355) [this resulted in some nice pictures](https://youtu.be/Pnrizu4LKO8?t=358) [including one with the Bear in bed with](https://youtu.be/Pnrizu4LKO8?t=363) [the night cap on](https://youtu.be/Pnrizu4LKO8?t=367) [this is what I wanted](https://youtu.be/Pnrizu4LKO8?t=369) [so that's all I wanted to share with you](https://youtu.be/Pnrizu4LKO8?t=373) [today I hope this gave you some](https://youtu.be/Pnrizu4LKO8?t=374) [inspiration to go and generate some](https://youtu.be/Pnrizu4LKO8?t=376) [images of your own](https://youtu.be/Pnrizu4LKO8?t=378) [please share your Creative Image](https://youtu.be/Pnrizu4LKO8?t=380) [generation workflows and the links to](https://youtu.be/Pnrizu4LKO8?t=382) [the pictures you've created in the](https://youtu.be/Pnrizu4LKO8?t=385) [comments below](https://youtu.be/Pnrizu4LKO8?t=387) [thank you](https://youtu.be/Pnrizu4LKO8?t=389)