Inspiration
Marine debris is harmful to underwater environments. Developing a machine learning algorithm that can identify plastic debris can be useful in creating more sustainable underwater environments for healthier oceans.
What it does
Displays images of various objects that may or may not be plastic.
How we built it
Using online open source data from the TACO dataset we were able to obtain images of plastic.
Challenges we ran into
Locating datasets of plastic debris underwater. Understanding how to read and utilize JSON files. Trouble preprocessing images to use for training. Trouble implementing VGG16 using TensorFlow.
Accomplishments that we're proud of
Displaying images.
What we learned
Different use cases for different Convolutional Neural Networks. That we have much to learn about using Deep Learning Models.
What's next for Machine Learning for Plastic Marine Debris
Using the dataset we have to train a model.
Built With
- google-colab
- python
- tensorflow
Log in or sign up for Devpost to join the conversation.