Finally. How we can automate colony counting without todays limitations

Learn how artificial intelligence shapes microbiology, with accurate automated counting and characterization of bacterial colonies. Including applied use cases.

Before investing in automated colony counters, read this startling study

In today’s automation of procedures in microbiology, laboratory technicians still struggle to make the process of colony counting easier. They want to achieve more confidence in the accuracy of their total colony count and Colony Forming Unit (CFU) calculations.

Due to app colony counters being surprisingly limited, and professional lab equipment often outdated, automation of highly accurate colony counting has so far been out of reach.

Challenge

In this drinking water use case, the company wanted to automate bacteria colony counting with high accuracy and high throughput, while distinguishing different kinds of bacteria. The challenges;

  • Counting each colony as one, identifying single clusters of colonies.
  • Accurate analysis of colonies on plates is difficult by the presence of dust or scratches.
  • Extracting other variables, such as size and color, to identify different kinds of bacteria
  • Gaining reproducible results.

Solution

AI provides a breakthrough. You will learn how;

  • Deep learning technology learns to recognize and distinguish different kinds of bacteria.
  • Pre-trained AI-powered colony counting models allow microbiologists to count real-time. 
  • Highly accurate, reproducible results, distinguishing colonies from anything else on the Petri dish, are possible.

Please enter your email address and a PDF copy will be sent to your inbox straight away. 

USE CASE DRINKING WATER COMPANY

Learn from Water Analyses

DATA CAPTURE

How to capture digital images of bacterial colonies on petri dishes

ANALYZE & COUNT

Learn how the AI system analyzes and counts overlapping colonies

TRAIN & VALIDATE

See how the microbiologist validates and corrects the system

OPTIMIZE & DEPLOY

Learn why they build the business case to integrate the system in the lab

1. Connecting image sources

Connect your laboratory devices, microscopes or quality inspection cameras to AI software, or directly upload plate images to start analyzing and counting colonies. 

  • Set the right lightning conditions to capture images
  • Scan barcodes of the petri dishes that will be captured
  • Keep uploading new images to continuously learn more

2. Analyzing bacteria colonies

Powered by deep learning, the system starts analyzing images of bacteria colonies. The software gives initial predictions based on a powerful counting algorithm, encircling every single colony.

  • Analyzed an initial dataset of 50 images of bacteria colonies
  • Zooming in on images to identify even the smallest colonies
  • Deep learning learnt to distinguish dust and scratches

3. Validating predictions

The predictions of colonies are presented to a microbiologist-in-the-loop to either correct or approve them. It learns quickly through real-time user feedback.

  • Indicating if the predicted bacteria colonies are valid or invalid
  • Correct invalid annotations with pen, erasing or marking colonies
  • Saving adjusted annotations to quickly improve the model

4. Optimization to deply

The more annotations and validations of the images are provided, the more advanced the model becomes. The predictions improve through interactive learning until deployment in real-time.

  • The system recognizes different kinds of bacteria
  • Counting numerous colonies in a matter of minutes
  • Providing real-time reproducible results in colony counting

Get your full report about
AI-powered colony counting

Understand how to obtain reproducable inspection results, taking errors out of manual and conventional counting systems.

  • Distinguish different types of bacteria colonies
  • Automatically recognize bacteria by its morphology
  • Retain and improve accuracy levels at rising volumes

Enter your email address and a copy of your PDF will be send to your inbox straight away.