-
TreeVis
-
Interface
For the average student, although we all want to have good sustainability practices, large amounts of information to know as well as varying waste-disposal standards create friction between us students and playing the part we want to in creating a more sustainable tomorrow. Having a friend by your side to ask throughout your day to day, whether you're cleaning up after your dining hall scraps or unwrapping packaging of different materials, would clear up barriers and allow ease-of-access to information.
Welcome to TreeVis, a JARVIS-style voice assistant carefully catered and designed to fit your sustainability needs.
TreeVis utilizes a FastAPI pipeline to a 2D U-Net implementation to scan surrounding objects for waste classification upon request, allowing users to correctly dispose of waste when recyclability and compostability are available. Speech recognition and segmentation classification using Google's SR API and OpenAPI, respectively.
Ethically, TreeVis allows for significantly better access to sustainable practices. The application does heavily utilize ML/DL models, but the implementation either heavily restricts tokens to minimize environmental impact or runs locally.
When it comes to sustainability, TreeVis acts as an immediately deployable, easy-to-use service for students at the 5C's as well as many other universities to promote eco-friendly practices in their day to day.
Log in or sign up for Devpost to join the conversation.