/wade-ai

an AI algorithm for detecting trash in geolocated images

Primary LanguageJupyter Notebook

AI tool WADE

Let's Do It Foundation (LDIF) is working on building an AI solution capable of detecting trash in images and videos. Video of the project is here https://youtu.be/zsgpqsRRzTE

Approach

The solution was build using Mask R-CNN and trained on dataset consisting of almost 1400 crowd sourced images containing trash. Ideally we would end up with a computer vision algorithm that could detect trash from any picture or video that has a geotag attached to it.

Team

Currently LDIF AI team is consisting of 7 people but we are always looking for more data scientists to help out as much as they can. All of the software that we create will be open source and might have other applications as well. Since our datasets are currently pretty small our main challenge is to create an algorithm that can be trained with the least amount of training data.

Get in touch

Feel free to contact us with your questions and proposals:

Kristiina Kerge
Technology Manager
kristiina@letsdoitworld.org