The project
noze applies machine learning techniques to the analysis of polysomnography and actigraphy data — the examinations that monitor sleep through physiological signals (brain activity, breathing, eye movements, muscle activity) and the activity-rest cycle.
The technical approach
The work focuses on:
- Time series analysis: processing and classification of complex physiological signals acquired during sleep. Machine learning algorithms analyse temporal patterns to identify sleep stages and anomalies
- Wake-up anticipation and prediction: predictive models capable of anticipating the moment of awakening, with applications in improving sleep quality and managing related disorders
- Path towards EU Level I medical certification: the system is designed from the outset to meet European regulatory requirements for medical devices
The startup
The work leads to the founding of the startup SleepActa, registered on 5 January 2017. SleepActa is established as a dedicated entity for the development and commercialisation of artificial intelligence-based solutions for sleep analysis.
The context
SleepActa represents the convergence of two noze trajectories: the machine learning expertise built over the years and the interest in digital health. Sleep becomes the application domain where these skills translate into a product with medical certification potential.