Another Kind of Self-discovery

Another Kind of Self-discovery

#BrainsIn the past weeks my creative juices have been quite active again. I stumble over things by accident and ideas and „what-ifs“ seem to take over my mind with a tight grip. When time allows it, I indulge in them and it’s quite interesting to see what happens. For example, this week, I found some old MRI scans of my head. Don’t ask how I got those back in 2015 – it was a mixture of curiosity and persistence, as it is always the case with me. But when I saw them these days, I wondered: What can I do with these? Can I take the information they contain and make something new out of it? Shouldn’t I be able to generate a 3D model with the help of some clever algorithms? The game was on. There wasn’t a purpose or application. And there still isn’t. There are just possibilities. For another kind of self-discovery.

The Source Material

These scans where what I had. And an idea that grew into a plan.

I had 7 of these scans, each containing 20 to 24 photos. These slices should theoretically be sufficient enough to render a 3D model if arranged correctly. That was the theory. Photoshops function to create so called 3D volume meshes should be they key to my plans. It uses greyscale pixel information of different photos arranged next to each other to interpolate the values between those photos and create 3D models out of them. It turned out that my source material was crap. 20 to 24 photos are simply not enough to depict what lies between them on the scale of the human head. Especially not if what lies between them is the most complex structure known in the universe.

Information Optimization

Seeing the first results, I realized: I’d have to support the algorithm of Photoshop’s 3D engine and deliberately reduce the information to make it easier to interpolate at least the most important structures of the brain. So I refined the photos and the results were quite nice:

Optimizing the photos included the following adjustments:

  • increase contrast to 100%
  • increase clarity to 75%
  • dehaze around 50%
  • sharpen 100%
  • noise reduction set to roughly 50%
  • Full 3D Renderings (Standard View)

    Three of the seven views were interesting for me to render – the lateral view from above, one frontal view and one superior view. The results are these 3D models that could be moved in space as desired.

    Though the best perspective was indeed the one they were taken from. Unfortunately I was not able to combine all three to one detailed model, Photoshop seems to have some limitations here. If anybody could help me here, I’d deeply appreciate it 😉

    Retouched Images & Partial 3D Renderings

    When I saw these, I felt I wasn’t quite finished. I wanted to go deeper. Much of what I saw here was skin and flesh and bone, blurring the interesting stuff inside this funny shaped egg-head. Everything that makes me … me. So how could I get to the core of it all? The answer was quite simple: work on every single slice and cut out everything that I did not want to be taken into calculation for the 3D model. So I masked every photo. Yay -.-

    Put THOSE slices into the 3D engine and you get this awesome model of your deepest core:

    Seeing these models, I was pretty satisfied what was possible by just using those crappy photos, working on them and using some Photoshop algorithm magic. But of course, just playing with these models in space wasn’t enough. I wanted to show the results and make something cool with them. I wanted to create some animations with these. The kind you see in Sci-Fi movies and futuristic observation interfaces.

    Animations of Rotating 3D Models

    Again, I did not have to use another tool but Photoshop. In Photoshop you can create a timeline and define the frames of these animations step by step. Of course, if you wanted to create 8 seconds of animated video, you’d have to do that 240 times (8 seconds multiplied by 30 frames per second which is typical in order to get smooth movements). Fortunately, animations based on 3D models just need some key frames. Photoshop does the rest. So I set these key frames to 2 seconds, 4 seconds, 6 seconds and 8 seconds making up a simple 360° rotation: 0 seconds standing for 0° or 360°, 2 seconds for 90°, 4 seconds for 180°, 6 seconds for 270° and 8 seconds for 359°. The result was amazing and combined, I got a nice dashboard for my animated MRI 3D models:

    So … this was fun. A creative way to discover some more of Photoshop’s magic aka awesome algorithms. To learn some cool stuff about the human brain – differences between cerebrum and cerebellum for example or what the heck all these ventricles are you read about in every article and diagram about the brain. And of course some reflection on the big questions: What makes us human? What makes us ourselves? What is the key to all the mess we as humans have brought upon this world? And to all the miracles and those to come? In my opinion many of us think they know the answers. This animation shows it pretty well: it is our brain, the most complex structure in the known universe. But change the perspective just a little, try to go deeper and you’ll see: there are still many gaps in our knowledge. We still have much to learn. I leave you with these questions and thoughts and my golden rule: be constantly curious and out of curiosity, create something good!

    Kommentar hinterlassen

    Deine E-Mail-Adresse wird nicht veröffentlicht. Erforderliche Felder sind mit * markiert.