bionomous logo

The EggSorter allows to automatically observe, analyze and sort out small biological entities, in particular the eggs of zebrafish. It is an affordable solution for laboratories working intensively with zebrafish, and soon for other biological research fields. The device does not even require an external computer, as all the processing is embedded, and the user can simply interact and configure the device via a convenient tactile screen and Android application. The device is portable and thus can be easily transported anywhere in the laboratory.


The core of the EggSorter device is a rotary wheel that can grab the eggs one by one and position them in front of an optic for inspection. The picture of the egg is then retrieved on an embedded computer, saved on the cloud and processed using various algorithms based on machine learning to classify each individual egg. After this process, each egg can either be maintained inside the device for continuous inspection, or it can be evacuated and placed inside multiple containers, according to its categorization.

Technology of the eggsorter: a wheel grabs zebrafish eggs/embryos one by one and places them in front of a camera. Deep Learning Algorithms analyzes the images and classifies the eggs, sorting and dispensing them based on that classification.   


Specific algorithms are deployed on the device to be capable of sorting the eggs according to several criteria, such as:

  • Fertilized or not-fertilized
  • Marked or not-marked (with Green Fluorescent Protein (GFP) for instance)
  • Development stage
  • Anomalies present during the development stage

This algorithms can also be adapted by the user, as they will depend on the types of research the user is performing with the eggs.

GMM example


We developed convenient user interfaces for the users to easily interact with the device and benefit from the maximum of functionalities that are offered. The EggSorter is controlled via an Android app that is easily parametrized. The App runs on an embedded computer and the user can interact via a tactile screen. Larger screens can be plugged in order to project the pictures of the eggs on a high definition screen.



Egg processing time1 sample/second
Size range of the entities0.6 to 1.4 mm diameter
Zebrafish egg stage fertilization determinationFrom 4 to 8 hpf
Output compartmentsFalcon tubes
Multiwell Plates
Petri dishes
Max. nb. samples for continuous observation30 samples


Dispensing time2 sec / sample
Multiwell Plate external dimensions85 x 128 x 15 mm
Petri dish external dimensions92 x 16 mm
Water dosing resolution30 μl
Minimal dispensing volume50 μl
Zebrafish egg survival rate>98.5%


False discovery rate4%
Wasting of fertilized eggs<15%


Weight15 kg
Length44 cm
Height51 cm
Width40 cm