MOODBOARDS WITHOUT ANY EFFORT
Endlessly searching for images to illustrate your ideas during a creative brainstorm?
It’s a useful but time-consuming task. Students from the Applied Computer Science program at AP University College developed a smart moodboard app to simplify this image hunt during the ideation phase.
Computer science students Guning Deng, Andreas Huygelen, Jasper Matheussen, Anne Scheyltjens, and Niels Van De Ginste developed a moodboard assistant, supported by various AI tools. The AI-assisted moodboard suggests visuals based on previously uploaded user images and conversations with the integrated chatbot. This allows you to visually shape your ideas in a single, unified environment.
Chatbot with a Dutch Accent
For the integrated chatbot, the team used the Vicuna 1.5 7B Large Language Model. Among the available variants, they selected Vicuna 7B—a version that: can be used for commercial purposes,requires only moderate computing power, and the Dutch language fluently and naturally.
Stable Diffusion
To generate images, the students opted for Stable Diffusion—a popular open-source diffusion model that receives regular updates and allows users to generate unlimited images for free.
Want to try this clever student experiment yourself? A live demo is scheduled during the coaching session on 28 March 2024!