Artificial Intelligence Archives • Verhaert Masters in Innovation

Computer vision algorithm repository

Nicky Sterck — Mon, 20 Sep 2021 15:45:46 +0000

We have a vast library of both deep learning and classical computer vision algorithms that can be easily optimized and adapted for your use case. This includes detection and classification, tracking, segmentation, anomaly detection, tracking, image registration, and many more.

The post Computer vision algorithm repository appeared first on Verhaert Masters in Innovation.

Privacy preserving people monitoring

Tess Loete — Mon, 16 Aug 2021 09:32:22 +0000

Contactless identification technologies such as facial recognition have boosted tremendously during the pandemic. But what about people’s privacy? Our computer vision technology detects masks, estimates distances, detects and counts people. Because the data is processed locally, no image ever leaves the sensor.

The post Privacy preserving people monitoring appeared first on Verhaert Masters in Innovation.

Chronic heart rate failure prediction

Tess Loete — Mon, 07 Jun 2021 09:11:54 +0000

This multi-sensory AI system (based on machine learning) analyzes measurement data and compares it with personal threshold values. With this information it can generate an automatic alarm where necessary so that the treatment can be adjusted based on the current state of health.

The post Chronic heart rate failure prediction appeared first on Verhaert Masters in Innovation.

ML for oil degradation

Tess Loete — Mon, 31 May 2021 12:00:19 +0000

The oil degraded working of gearboxes is characterized by different vibrations when compared to nominal behavior. The ML we developed is built to detect the difference between both states very accurately.

The post ML for oil degradation appeared first on Verhaert Masters in Innovation.

Context-aware system

Nicky Sterck — Fri, 21 May 2021 10:04:25 +0000

Detecting, localizing and measuring distance of objects, posture, people, vessels, and many more. This technology is LIDAR or RGB-D-based and segments objects out of the point cloud. Special model-based algorithms focus on capturing movement dynamics.

The post Context-aware system appeared first on Verhaert Masters in Innovation.

ML for bearing error detection

Tess Loete — Wed, 10 Feb 2021 11:39:43 +0000

By classifying samples (data blocks) from a data set with a bearing errors (BEAF), we set up an ML where the probabilities for all classes are plotted. Based on this information we can detected several types of errors and improve maintenance of these structures.

The post ML for bearing error detection appeared first on Verhaert Masters in Innovation.

ML for grinding feeds

Tess Loete — Tue, 01 Dec 2020 13:04:22 +0000

We designed a grinding feed ML to ensure grinding wheels could last longer and could not be used once worn out. This feed was based on data from three 3D accelerometers and an ML with the appropriate vector functions to distinct between two states.

The post ML for grinding feeds appeared first on Verhaert Masters in Innovation.

Fire propagation prediction

Tess Loete — Tue, 06 Oct 2020 14:25:52 +0000

For fire prediction, propagation within a phase and evolution from one to another is much more important than detecting the stage of the fire. An ML with several classes can determine the speed of propagation from one state to another.

The post Fire propagation prediction appeared first on Verhaert Masters in Innovation.