Gangen Alumni, at its surface, is a digital art NFT collection of 10,000 characters — with unique art, a suitable narrative, and a pragmatic roadmap.

In this entry, we’ll explore the concepts encapsulated by the collection and detail what went into the production of the NFT characters.

The Art

Let us first discuss the current standard of NFT art and why changing methodology is important to keeping the NFT art market fresh.

It’s become quite clear that the simplest and most commonly used approach to producing NFT digital art is through the use of a fixed model with interchangeable attributes. It entails using a single model as the base to a character or art piece and then applying layers of designed components onto the model. These components, in most circumstances, are independently re-used between each art piece, however, changing the combination of components used within a given art piece makes it a unique art piece within its collection.

Despite many projects genuinely leveraging this approach to serve amazing unique works, some even appearing on NYC billboards, some incorporating animations, and other differentiating facets, this approach to NFT art has given rise to mass-produced art for the sake of the sale.

Not all collections follow this approach though.

Pixel art is an approach to procedurally generating great visual experiences and is used in the most prominent collection to date, as well as in many major video games.

Collections have also taken inspiration from existing 2D collections and 3D-ized them, whether it be 3D pixel art or 3D model-based art.

I used an NFT art collection as a catalyst to learn how to develop for Crypto, specifically to develop Solidity Smart Contracts. Since digital art is imperative to an NFT collection, I was struck with a quandary. Shall I fast-forward the digital art process, and treat this project as an educational milestone, or genuinely attempt my unique take on digital art?

I chose the latter and decided that I was going to take an artistic approach that I’ve not seen anywhere else.

Being fascinated by AI technology and aware of generative image production with AI, it became quickly apparent that procedurally generated AI art could be a captivating approach to an NFT collection.

The criteria were quite simple: break out of the mould and structure of a fixed model.

This endeavor proved difficult and lengthy, taking many months to reach a point of artistic satisfaction.

I began with a foundational AI-generated layer for each character.

The produced foundational layer of each character proved to deliver on the criteria, however, resembled real-world humans too much to be considered art. It was almost photography. Despite enjoying the many faces of humanity in this preliminary collection, it was quite underwhelming to see profiles of human-ish characters when you could step out into the real world to see the very same thing in higher definition.

Considering this, a subsequent explicit decision was made to process the generated foundational layer through AI technology for filtering human faces into Anime/Manga faces.

As an admirer of Anime and the absurd and hyperbolic worlds that can be created due to this unrestricted animation style, I immediately found narrative direction in this art style.

Additional AI processing to heighten the colour spectrum and refine the characters into, what could be considered, “drawings” — and we have our final result.

The result truly fulfilled the criteria I had planned for the collection and much more.

Each character art piece of the Gangen Alumni collection:

• evokes unique emotion
• can have a different complexion, resemble a different race and gender
• has a unique hairstyle or headwear
• can face in different directions
• can wear different colours and apparel
• be presented in front of different backgrounds
• can be accessorised differently — taking some inspiration from the fixed model art approach

The Technology

With deeper expertise in software engineering than in illustrative design, exploring the use of a GAN (Generative Adversarial Network) as the foundation to the art pieces in the collection was pivotal in deciding what additional layers of AI could be applied to yield a unique art style for the collection.

A GAN is effectively an AI model for processing many images (considered training data) of different objects, animals, humans, or even animated characters, and then producing a unique rendition.

The GAN only produced the foundational layer of the art piece. The aforementioned Anime filtering was then used to characterise and animate the produced humans.

Anime is a very well explored category of animation due to its popularity across the world. There are quite a few very well-documented explorations of Anime and AI. There are even companies using Anime focused GANs to dynamically produce new characters for video game companies, or Anime studios.

During development, I decided to explore OpenCV for supporting facial recognition, however, the cascade files provided by the OpenCV library only supported human faces.

It was until I stumbled upon this library that I realised how much effort is being put into Anime related AI efforts, and just how impressive they are.

Beyond AI face filtering, static accessorisation seemed necessary to continue distinguishing characters. Facial landmark detection played a huge role in this effort, as it provided a mechanism for applying static images to appropriate parts of the face, and doing so relatively successfully across many different facial sizes and poses.

A facet of an NFT collection that is often undervalued is NFT metadata.

This categorisation and facet management is imperative to enabling a narrative to each art piece as well as providing an interface for filtering art pieces when they’re listed on a public exchange.

The problem I faced in producing this metadata sufficiently is that our artistic approach starts with an image. Traditionally an artist starts with the facets that make up an image to produce the image.

For this reason, we turned to facial analysis technology to extrapolate metadata from the AI-produced images. This included detection of gender, probable age, emotion and more.

AI vision and image labeling were used to further improve metadata richness, detecting what type of hat a character is wearing or what could potentially be in the background.

AI is probabilistic in nature, so all detections and analyses yielded confidence scores. Additional software was required to parse out and apply detections that met a given confidence threshold.

Developing this technology was highly experimental and was contingent on the outcome of the produced image.

Nonetheless, the technology and combination of different experiments and AI layers was my art. The produced images and rich metadata are simply a representation of where my passion was applied.

The Narrative

Who are these 10,000 1-of-1 characters without a world for them to exist in?

“Metaverse” is a phrase used across many collections to establish this sense of “world”, however, it is important to understand what a Metaverse is in order to determine whether the world of Gangen is one.

“The Metaverse” is commonly understood to be this virtual landscape where humans can connect without language or geographic constraints, where digital avatars and characters will become our identity.

This notion seems far-fetched but is viable at a point in time when there is an economic incentive to exist within a virtual landscape that out-incentivises anything offered in the physical world.

This is why some individuals with highly regarded opinions reference “The Metaverse” as this point in time.

The problem with this notion is that the current landscape of NFTs is fragmented. Each collection that establishes a sense of “world”, is its own Metaverse, resulting in a landscape of “Metaverses”.

This fragmentation takes power away from this concept of “The Metaverse”, and therefore this label is something that I do not think applies to this collection.

Gangen Alumni is not a world you embed yourself in, nor does it play a role in creating a virtual micro-economy that requires an investment of time for earnings that out-perform opportunities in the real world. This is not a play-to-earn game.

There are perks and benefits that are earned through roadmap activations and there are economic incentives to participate in the ownership of the collection, however, Gangen Alumni is architected for the entertainment and engagement of an audience — making it more like a Show, or a Sitcom.

Ownership plays a role in this collection in the same way owning the rights to a Movie Script or being an agent for an Actor or Actress can yield economic opportunities, but the agents and owners of a show do not facilitate or make up the world they have a stake in.

The show is its own autonomous fictional artistic expression.

For this reason, I consider Gangen Alumni a meta-Show.

Becoming an Agent

Owning and taking agency over a character yields access to exclusive benefits and perks which can be unlocked by the community through roadmap activation.

As it is a Show, there is a balance in these unlocks between delivering art and entertainment and embedding economic opportunity for Agents.

One such example of this is in our planned second collection of Character Backstories. This collection will be airdropped directly to Agents. These stories are not planned to fit within the structure of a simple OpenSea NFT listing. The stories are text-based, leaving the section for media open to engagement by the community. There are interesting ideas planned for how to allow the community to engage with this Show, in a way that resembles how businesses or individuals engage with network TV shows.

Fundamentally, a community engaged with this collection will drive our efforts to deliver on the roadmap activations and deliver benefits for Agents.

Take your first step in owning a stake of Gangen Alumni.
Purchase the rights to a Character of the NFT collection today. Each Certificate of Character Rights can be used to mint a Character during the NFT collection sale.

Interested in learning more?

Visit the Gangen Alumni website

Join us on Discord and follow the Show on Twitter and Instagram for updates.