Neural Style Transfer: A Novel Style Attribution Framework

Neural Style Transfer Attribution at work. The algorithm has been improving very specific cases over time, but it’s a work in progress.

I did a deep dive a few weeks back on Neural Style Transfer — a fascinating AI technique that allows extraction of the style of an image and the application of that style to another image — at the Flatiron School for the AI-Python Meetup for a Kinder NYC.

While this is an amazing technology, being used in image generators all over the world now, we must acknowledge the impact that such tools have on artists. The style of a piece, for many, represents their life’s work and the culmination of years and years of hard-earned skill to achieve.

For those who haven’t opted in — To have their efforts be stripped and then applied to another piece of media at an instant, and then distributed often for profit, without any compensation to said artists, is not only unethical, but unkind and irreparably damaging to the spirit of human creativity.

Here is the video of the outline to build this project we did from building front-end to ideating and composing the M.L backend for the Attribution Function.

Thank you for reading and watching!

Please follow me on the following platforms for more updates:

Linked In: https://www.linkedin.com/in/shafik-quoraishee/

Youtube: https://www.youtube.com/@shafthealgoguy