Home Sleep Study



Clebre is a Semantive’s start-up: a home sleep study to diagnose apnea or snoring. Breathing problems are common, and they are also commonly undiagnosed. For example, 10% of children are habitual snorers; therefore thousands of them may potentially require surgery because children who snore are at higher risk of neurobehavioral or physical abnormal development. Nowadays, the diagnosis is based on the history obtained from caregivers, and rarely on an unbiased study like in-lab polysomnography. Such a full sleep study is expensive and limited, which is why children are not correctly diagnosed or treated.

Clebre wants to change it. Clebre delivers a device for home testing of a breathing disorder, which is cheap and more comfortable for patients than in-lab polysomnography and is also easily repeatable.

Our work

  • a wireless sensor prototype;
  • algorithms for signal analysis;
  • an Android application;
  • a server for data collection and mass analysis.

The outcome

The prototype of a home sleep study device. The algorithm we designed includes the signal filtration, segmentation to extract consistent episodes, and events classification with a combination of ANNs. Our solution holds a patent granted by the European Patent Office.

Technology used
  • Python
  • Convolutional Neural Networks
  • TensorFlow
  • Keras
  • Horovod