A Camera for the Mind: The Advent of Neurography

Update: read our new post going into more details about this research and topic:

Have AI Brain to Image Systems Arrived? Not quite.

---

A ‘camera for the mind’ is one of the most compelling analogies to describe AI image generation. We translate our imagination into words, then AI translates those words into images.

Where photographs use photons to capture and recreate moments of reality, is AI image generation doing the same with neurons and moments of imagination? Is this neurography? Are these neurographs?

Not quite, there are too many layers of abstraction between the neurons, thoughts, words and the resulting images for this to be a wholly accurate neologism. Although it remains a good analogy.

The true advent of neurography would require thoughts to be the prompt, rather than words and it seems we’re getting closer to this future.

On November 21st 2022 a paper was released by Google and Osaka University titled ‘High-resolution image reconstruction with latent diffusion models from human brain activity’ showed the ability to recreate what someone is seeing with their eyes.

The abstract reads:

"Reconstructing visual experiences from human brain activity offers a unique way to understand how the brain represents the world, and to interpret the connection between computer vision models and our visual system.

‘Understanding internal process of Stable Diffusion with encoding models of brain activity’

It could be argued the age of neurography may be in its early days. (We’ll be explaining more about what this paper does and doesn’t mean soon.) With recent advances in AI image generation, recreating what someone is seeing in their minds eye feels one step closer.

While initial reactions mimicked those to Louis Daguerre’s introduction of photography, which Nadar recalled being:

“so utterly unforeseen, so far beyond anything imaginable, it destabilized all that was known and believed.”

They were pre-mature and this technology, while amazing, isn’t so new. Its marriage to stable diffusion AI image generation is an exciting step forward however.

Peyman 𝕄𝕀𝕃𝔸ℕ𝔽𝔸ℝ @docmilanfar

So much overhyping, it makes your head spin Brainwaves to images is something that’s been around since at least 2008. To somehow suggest that hooking up stable diffusion means that it is “literally reading people’s minds” is absolutely ridiculous. scientificamerican.com/article/transl…

deiniolb 🐻👉 aisuite.io @danberridge

I'm speechless. Not peer-reviewed yet but a submitted paper. The 'presented images' were shown to a group of humans. The 'reconstructed images' were the result of an fMRI output to Stable Diffusion. In other words, #stablediffusion literally read people's minds. Source 👇 https://t.co/sne2LhF8wZ

As we said in our inaugural post about photography and AI image generation: We're there again. Enamoured again. Fascinated, excited and scared again - of a new power, from a new prometheus - given to us: mere mortals.

If the neurographic age hasn’t begun, it is at least in the offing.

---

NEUROGRAPH | nəˈtɒɡrəfi | 

Noun • an image made using neural activity as a prompt for an AI imagine generation system

---