Cases and projects

Image recognition that counts animals and tells species apart on wildlife camera images. MIT Zuid feasibility project from 2018, continued in the population dynamics platform with WUR, where species are recognised with an accuracy above 98 per cent.

In the MIT R&D project AI-based medical assistant (2018 to 2020), Twentynext and WSK Medical researched image recognition that finds cells and nuclei in digital biopsies and classifies protein expression.

With OMC Zaandam and MagnaVersum, supported by the Dutch MIT scheme, Twentynext researched how machine learning helps ophthalmologists recognise abnormalities in eye images early.

With the veterinarians of seal rescue centre A Seal: an AI system that recognises lungworm infections in seals at rescue centres and determines the stage.

Skully Care measures skull deformation in babies from a single photo. Twentynext develops the algorithm that automates that measurement.

Asthma patients are grouped by data analysis into segments with their own characteristics, so treatment and medication can be tailored to them.

A medical device with body sensors and controlled drug delivery, connected to a Clinical Decision Support platform that clinicians monitor remotely.

For researchers at Wageningen University we are building a non-invasive monitoring system around the critical transitions that test resilience.

For AxionContinu, a nursing, residential and home care provider in the Utrecht region: reports and dashboards for task and management information from client records, rosters and finance, with the same definitions. Plus AI applications that support care professionals. In use.

Subversive crime leaves traces, but rarely in one place. For DITSS we built a crime room where scattered signals come together, so analysts can interpret them sooner. Developed from 2018 to 2020, handed over to the client in 2020.

Within Q-LAB, the Dutch National Police innovation programme, we examined which combinations of existing sources yield new information and which mainly produce spurious correlations. Not a platform, but a substantiated answer to where linking pays off. Since 2020.