FURI | Spring 2022

Creative Frameworks: Leveraging Deep Learning and Data Analysis to Create Accessible Artistic Technologies

Data icon, disabled. Four grey bars arranged like a vertical bar chart.

Generative adversarial networks are used to create visuals from textual inputs, allowing a user to create artwork with only a keyboard. GANs can be extended by recycling the output image as a seed for another generation, allowing the creation of dynamic videos. Using this technology, individuals can create a video by simply entering a list of image prompts. Combining this generative image technology with auditory analysis and Digital Multiplex Control, researchers have made creative expression across digital mediums accessible to users of all backgrounds and abilities, removing traditional limitations of technology/specialized software access and complex physical interfaces.

Student researcher

Joshua Pardhe

Joshua Pardhe

Computer systems engineering

Hometown: Phoenix, Arizona, United States

Graduation date: Spring 2022