With the rise of NFTs (Non-fungible Token) in the world of digital art, many collectors around the world start relying on this digital ledger without actually understanding all the risks, one of which being the authenticity of the copy. Especially, with the plethora of vulnerabilities shaking off the concept of ‘authenticity’, one of the industries that have been most-affected by this is digital art. Thus, we have decided to apply blockchain technology to solve the authenticity issue in digital art.
What it does
ARTistory has two main functionalities. (1) The original creator of an art comes onto our ARTistory web app and registers their work before selling it off via NFTs. (2) Our blockchain will preserve the authenticity of this specific artwork, so a questioning buyer can check whether they have got the real copy of the artwork on our web app.
How we built it
We have used Node.js framework for the base server that handles registration and checking. When received an uploaded artwork, the web app will create a new block with a checksum that is based on all metadata from this artwork and mine to make a new contract in the blockchain, which will then be broadcast via a P2P (peer-to-peer) server. This artwork will also be uploaded to MongoDB Atlas, which will then provide the sources for our second functionality that uses DeepAI to identify how similar the image that gets checked on is. We have deployed ARTistory using Google App Engine on Google Cloud Platform (GCP).
Challenges we ran into
We have come across three challenges during this project. The first two come through the development process. Building the blockchain and calculating checksum that takes metadata into account have taken us some time because this is our first time implementing a blockchain. The last challenge was the biggest one and we came across it during the deployment. The duration of the DeepAI image similarity checking depends on how many artworks have been registered, and our deployment service provide, Heroku, has a default request time-out of 30 seconds, so, if there are too many registrations, the checking functionality gets too slow or does not work at all.
Accomplishments that we're proud of
One thing we are proud of, regarding this project, is that our idea can become a potential contribution in solving the issues with artwork copyright. Also, we were able to integrate a blockchain and similarity detection within 24 hours.
What we learned
We learned how to implement a blockchain and its inner working. We found out that there are many ways to calculate an image’s checksum and metadata and image bit to ensure the authenticity of the image. We also got a chance to explore the NFT issue.
What's next for ARTistory
This idea can be applied to other fields such as original image authenticity for combating fake news and image forgery since we also have the image similarity detection feature.