Iris, our automated colony counter, that works!

AI has become more and more important. We would like to tell you more about how we use artificial intelligence on our products.

If an analyst can recognize a colony type on an image within 1 seconds, a neural network can learn this too. This also applies to the distinction between particles, dirt, and colonies. Even the reflections of colonies are recognized as such. A neural network is therefore trained with the knowledge and experience of the analyst. Working with colony counters that are trained with an AI neural network is much faster.

The colony counter is based on artificial intelligence, which makes that it really works! Read more about artificial intelligence here. 

In our introductory course, we select two colony/micro-organisms combinations. We will train two neural networks on them. To adjust the camera and exposure, we will choose a number of representative petri dishes.

We begin by examining if there is an existing neural network available to support the initial annotation of the automated colony counter. These two configurations are stored as separate tests in the "TrainingSet Creator." Using the "TrainingSet Creator," you can generate and annotate images.

We oversee this annotation process and assess the quality, balance, and size of the dataset. Once the dataset is complete, we train a new neural network based on it. We then upload your newly trained neural network for evaluation in the "Validation app."

With the help of this “TrainingSet Creator” you create and annotate images.

The TrainingSet Creator is developed to create datasets needed for training neural networks. The photos can be taken in series, but it can also be done individually. If you choose to take pictures dish by dish, you can still use the petridish to assess while annotating. So you can annotate one picture at the time of annotate a series of images. 

This web application (intranet) is easily accessible through any web browser on your network. It works easily with a username and password. Because this is a web application, it does not matter where the validation is done. 

The Validation Web App is an intercompany web application that can be easily accessed via a web browser. You can log in with your username and password. Then choose which test or analysis you want to use, enter the Petri dish barcode by typing it or scanning the barcode.

Take one or more photos and validate or correct them. Indicate on the validation PC what the false positives and false negatives are and validate the image. Because this is an intercompany web application, it does not matter where the validation takes place.

After validation all information in a Zip file on the FTP server is made available for e.g. LIMS.

It is intended for small volume laboratories or as a replacement for traditional automatic colony counters. You have to manually handle the petri dishes.

The Iris counts colony's grown in petri dishes or petri film based on artificial intelligence.

Camera, A.I. server, FTP server (LIMS data) and Validation web app.

For 90 mm and 60 mm petri dishes and petri film.

It is currently under development. The 'Iris' desktop is all-round and flexible. It is intended for labs with medium volumes or wide variety of consumables. A robot takes care of scanning and loading the Petri dishes into the Iris. A maximum of 100 petri dishes can be loaded at a time, which are then processed by the 'Iris' desktop. Only in this variant, in addition to the 90 and 60 mm, also 150 mm petri dishes with 'twist lock' lids and petri film posibilities.

Robot, Camera, A.I. Server, FTP server (LIMS data) and the Validation Webapp.

It is intended for laboratories with large volumes of Petri dishes of size 90 mm. The production capacity of our colony counter Iris can be accelerated in combination with our system 'Irina'. The Irina platform with built-in Iris will: unstack the petri dishes, remove the lid, read the barcode, count colonies automatically and much more. Irina is a system that processes Petri dishes at a rapid speed for high throughput laboratories. It processes 360 Petri dishes in approximately per hour.

Irina platform, Camera, AI Server, FTP server (LIMS data) and The Validation Webapp. Only for 90 mm scales.