Detecting Microorganisms in Water Using Deep Learning

Local Chapter San Jose, USA Chapter

Coordinated byUnited States ,

Status: Ongoing

Project background.

United States does a very good job of providing clean and safe drinking water to most of its residents, but water borne diseases are becoming an increasing problem. According to the Centers for Disease Control (CDC), approximately 7.5 million waterborne illness occur annually, with a healthcare cost of about $3.3 billion. These infections result in emergency visits, hospitalizations, and deaths. These are caused by microorganisms, viruses, and fecal matter in the drinking water that are the result of an aging infrastructure, chlorine resistant pathogens, and finally an increase in recreational water use.

The problem.

Normally, water needs to be sent to a laboratory for tests. This is an expensive and time consuming process and the results are not immediately available at the point of use. What we need is an easily accessible and usable method for detecting microorganisms (in other words, bacteria) in drinking water.

Project goals.

- Create a low-cost method that is easy to access and easy to use for detecting microorganisms (bacteria) in drinking water.
- Train a CNN (or equivalent) to recognize and classify bacteria using Computer Vision techniques.
- Data: EMDS (Environmental Microorganism Image Dataset).
- Use inexpensive microscopes (paper/digital - with costs ranging from $20 - $100) of different magnifications (140x, 400x, 1000x) to find the best for the task.
- Deploy the trained and tested model on a mobile phone.

Project plan.

  • Week 1

    Research models/datasets. Summarize results (similar to a literary search).

  • Week 2

    Understand the data and if there is enough to fit NN. Discuss other methods – data augmentation, etc. Exploratory data analyses.

  • Week 3

    Train and test CNN. Train and test other models as well – Visual Transformers.

  • Week 4

    Choose the best model for deployment based on results.

  • Week 5

    Deploy on mobile phone.

  • Week 6

    Evaluate results for each microscope.

Share project on: