@vipyne

training data - aerial silk poses paired with stick figure images produced by running the aerial silk images through a pose detection script.

vipyne trained a model on a dataset of their own aerial silk poses paired with “stick figure” images produced by running the aerial silk images through a pose detection script. They later experimented with inputting drawings to the trained model to generate images.

showing the generated outputs from the model when given drawings as input.

Back to Participant Work Back to Participant Work