SUMMARY AND PURPOSE OF THE PROJECT
Hybrid InfraRed System for Affective Computing (HIRA) is a project funded by EU Eurostars.
HIRA will be an innovative facial imaging module for real-time affective computing, based on both behavioral and psychophysiological clues using advanced AI algorithms for super-resolution and emotional intelligence.
HIRA Driver Monitoring System TRL 9 release.
HIRA Pilot Monitoring System TRL 9 release.
Advanced algorithms leveraging dynamic information, sensor enhancing and fusion.
First hardware/software prototype release, compliant with automotive and avionics industry standards.
Conversion of the models in the TensorFlow lite format for deployment on edge devices available.
Validation study on the generalization of the super-resolution algorithms.
System architecture definition & preliminary version of the AI models.
Framework for Behavioral, Psychophysiological and dynamical AI.
Targets & hardware/software requirements.
Next2U is an innovative company that develops solutions for psychophysiological assessment, affective computing, and human-machine interaction. Leveraging non-contact, non-invasive technologies like thermal imaging, near infrared, and visible spectrum imaging, Next2U tailors solutions for diverse sectors such as automotive, avionics, social robotics, and simulated environments.
Next2U leads the HIRA projects offering expertise in human psychophysiology, computer vision and AI.
SUPSI-IDSIA is a pioneering research institute in deep learning for artificial vision, with expertise in autonomous robot navigation, medical imaging, and industrial process monitoring/control.
SUPSI, through the Swiss AI Lab IDSIA, is actively involved in the project, leveraging its expertise in machine learning, deep learning, image processing, optimization, and robotics.
TOELT LLC is a company that is actively engaged in multiple R&D projects, focusing on developing new technologies in the field of artificial intelligence, particularly in super resolution and the application of AI to sensors.
TOELT LLC excels in using advanced computational techniques to tackle complex scientific challenges, pioneering innovative methods for problem-solving.
HIRA fuses the spectra of near-infrared (NIR) and long-wave infrared (LWIR) radiation to compose a multi-spectral information set that gathers the behavioral and psychophysiological information to highlight the emergent patterns of the subject’s affective state.
The HIRA project utilizes artificial intelligence algorithms to extract significant insights from data streams in the NIR/LWIR spectra.
“It exploits innovative dynamic emotional intelligence to reconstruct the subject-specific affective state.”
Super resolution makes HIRA tiny and embeddable. Small, low-resolution sensors are capable of recovering the complete physiological and affective information from the subject.
The objective assessment of emotional and affective states in an everyday life context, needs integrated approach capitalizing on contactless assessment of both behavioural and psycho–physiological based clues.
HIRA uses affordable, miniaturized and low-resolution sensors combined with super- resolution algorithms to fill the gap with the more expensive hi-resolution sensors.
The classic and currently facial expression approach in use for the assessment of the subject’s emotional state does not have any chance to catch the same psycho–physiological clues as HIRA can do.