Holographic Microplastic Classification
A deployed CNN classifier for holographic microplastic images, paired with practical environmental context.
Overview · Implementation · Evidence
01THE MOTIVATION
The problem
Microscopic holographic images are difficult to classify by eye. This prototype investigates whether an image classifier and accessible demo can support exploration of polymer image categories.
STATUS / Working demo
02UNDER THE HOOD
The implementation
THE BUILDClassifies polymer types and dust variations from uploaded microplastic imagery using a PyTorch CNN and a Streamlit interface. The repository includes a public demo link and example image.
03ACTUAL ARTIFACTS
Inside the build
Screens from the application, with the workflow and context for each view.
04THE EVIDENCE
Read the source
Explore the code, run the application where available, or watch the original demonstration.