Inspiration
Saudi Arabia is known by its rich heritage. Therefore, we decide to show some of that.
What it does
We apply a AdanIN-Style transfer to apply the modern art to the abstract piece of art done by our artist.
How we built it
we use runwayml platform to train our model to use it in neural-style. We fine a pre-trained model trained on data set that represents the modern art. To address the difference in the data set used to train the pre-trained mode. We developed a script that crawl in domain images.
Challenges we ran into
Address the genre difference between the data set used to train the pre-trained model and the artist paint.
Accomplishments that we're proud of
The connection that we developed between the artist and the developer. Moreover, the we are proud of the work that we have done as it simulates the artist vision.
What we learned
We learned how to combine two different fields. Also, we learn how to leverage the art using AI.
What's next for futuristicArts
We are interested in adding more data that represents the in domain data and see how that could effect the finaloutput. Moreover, we will investigate the possibility of improving the Generative model.
Log in or sign up for Devpost to join the conversation.