Research

Articles

A selection of peer-reviewed work that supports ClarScan's technological direction.

Overview

ClarScan's technological direction is grounded in research showing that alcohol-related decline in functional capacity can be assessed using smart devices and machine learning. Studies do not treat alcohol solely as a chemical indicator in blood or exhaled air, but as a state that affects a person's movement, speech, reaction speed, fine motor skills, phone usage patterns, and eye-related biometric features.

Earlier work has shown that passive smartphone data, such as movement, device use, calls, typing, and time-based patterns, can help detect episodes of alcohol use. Active tasks have also been studied, where a person completes a short test: typing, pressing buttons, solving everyday phone-based activities, or reading complex phrases aloud. Such tasks matter because they measure a person's current state and functional capacity more directly.

Speech analysis studies show that alcohol can alter vocal acoustic properties, speech tempo, pauses, and articulation. Movement-based research links intoxication to changes in gait, balance, and driving behavior. Eye- and pupil-based work suggests that alcohol may also appear in patterns related to visual and pupillary responses.

ClarScan follows the shared conclusion of this body of work: the most practical path is not to rely on a single sensor, but to combine multiple data types into one multimodal risk assessment. Such a system does not need to replace a breathalyzer or laboratory test; it can serve as an early warning and decision-support tool in situations where a wrong decision can cause harm. For example in workplace safety, mobility and rental services, personal self-checks, or supervised use cases.

  1. Mariakakis et al. Drunk User Interfaces: Determining Blood Alcohol Level through Everyday Smartphone Tasks DOI: 10.1145/3173574.3173808
  2. Bae et al. Mobile phone sensors and supervised machine learning to identify alcohol use events in young adults: Implications for just-in-time adaptive interventions DOI: 10.1016/j.addbeh.2017.11.039
  3. Lee et al. S-ADL: Exploring Smartphone-based Activities of Daily Living to Detect Blood Alcohol Concentration in a Controlled Environment DOI: 10.1145/3613904.3642832
  4. Bone et al. Intoxicated speech detection: A fusion framework with speaker-normalized hierarchical functionals and GMM supervectors DOI: 10.1016/j.csl.2012.09.004
  5. Laptev et al. Combination of Audio Segmentation and Recurrent Neural Networks for Improved Alcohol Intoxication Detection in Speech Signals DOI: 10.3390/sym18020262
  6. Nassi et al. Virtual Breathalyzer: Towards the Detection of Intoxication Using Motion Sensors of Commercial Wearable Devices DOI: 10.3390/s22093580
  7. Segura et al. Advancing Intoxication Detection: A Smartwatch-Based Approach DOI: 10.1109/WF-IoT64238.2025.11270600
  8. Suffoletto et al. Detection of Alcohol Intoxication Using Voice Features: A Controlled Laboratory Study DOI: 10.15288/jsad.22-00375
  9. Tapia et al. Alcohol Consumption Detection from Periocular NIR Images Using Capsule Network DOI: 10.1109/ICPR56361.2022.9956573
  10. Thompson et al. Mobile Phone Sensor-based Nigerian Driving Dataset to Detect Alcohol-influenced Behaviours DOI: 10.1109/ICAC65379.2025.11196488

← Back to home