amelia gan - Emotive Canvas

Harvard University, Graduate School of Design
Enactive Design, Fall 2022
Instructor: Jose Luis Garcia del Castillo Lopez
Collaborator: Quoc Dang

Emotive Canvas aims to propose new methods for digital drawing by using the user's emotional state as parameters for controlling the tool. In this prototype, two modes of drawing are explored: one for use with the drawing app alone and the other for use in a multi-user configuration. When the app is used in single-user mode, facial expressions dictate both the line weight (brushstroke size) and brush color. This could be based on common emotion-color associations, such as red being associated with anger and blue with calmness. In multi-user mode, the line weight mapping remains the same, but color is used to assign a unique user ID.

The drawing app is a creative way to communicate without using direct video feed or text relays. It could be used as a mood or enthusiasm indicator in collaborative works, providing feedback through indirect cues. Other potential applications include a virtual whiteboard for brainstorming sessions.

Additionally, generative ML tools have been integrated into the canvas to provide ideation aid and interpretative prompts. The workflow includes using CLIP to convert the digital sketch into text and then back to an image again using Stable Diffusion. Parallel methods have also been tested usingOpenAI's Outpainting as a prompt for sketch completion or infilling.

happy brushstroke

angry brushstroke

neutral brushstroke

sad brushstroke

User 1

User 2

Embedding Suggestive AI