Who doesn't love dancing (at least in private)? There's an art to dancing and we wanted to show that art from a new perspective. The goal of this project was to combine the static art of painting with the dynamic joy that dance can express in an abstract and thought provoking way.
What it does
Art of Dance takes the beauty in movement and uses it as inspiration to create a digital color masterpiece personalized to the dancer.
How we built it
We used many different languages, each being particularly useful for a separate piece of our project. We got these to work in conjunction through data streams and server side storage. The difficult part was inventing and designing algorithms to convert the raw movement data into artistic images. We all contributed design ideas that compiled into several very different methods of displaying rhythmic movement as static art.
Challenges we ran into
- Parsing the sensor data using C++ into a format easily read in to Processing to generate the artistic representations.
- Accessing a pre-trained neural net API that could take our generated images and make them more artistic.
- A lot of accidentally overwriting each other's code on our github repo.
Accomplishments that we're proud of
- Reading raw movement data from multiple Myo devices to give a better perspective of actual movement of a dancing subject instead of interpolating movement from a single sensor.
- Organizing our tasks into modular code so that we could use tech tools to encourage collaboration in artistic algorithm development.
- Creating a more artistic image generation algorithm that procedurally used smoother color transitions to make an resulting artwork that seemed more professional and was easier on the eyes.
What we learned
- Functional possibilities of Myo Wristband technology.
- How to use technological tools to enable artistic collaboration, for example github repositories.
- How to use image generation tools like processing as well as neural network APIs to explore artistic possibilities.
What's next for Art of Dance
We want to look more into training a neural network to process our images out of a mostly geometric scope into a more creative and abstract artistic perspective of beauty of motion. We would also like to look into developing our own machine learning algorithm that can detect when a particular image generation method is colorful enough and saves that as a valid option for data representation.