Open innovation by SMK

Technology to Analyze Diseases
and Emotions through Speech

This technology enables analysis and diagnostic support for latent brain-related diseases and emotions through free speech in a few phrases of about 30 seconds. More specifically, it can provide diagnostic support for diseases such as dementia and depression for which early detection is important as well as analyze data such as stress, anxiety, and fatigue levels, which are difficult to detect quantitatively. Analysis is performed with any device equipped with a microphone and a network connection, and so is possible with widely-used smartphones, PCs and tablets.

Product Overview

  • 01

    Based on free speech in a few phrases of about 30 seconds, brain-related diseases and emotions can be analyzed.

  • 02

    By referring to 2,548 biomarkers contained in voice, highly accurate analysis at the level of medical devices is possible.

  • 03

    Analysis can be performed with apps on smartphones, PCs and tablets, etc.

Technical Overview and Product Specifications

Under Development

●Measuring Principle

It extracts and analyzes voice features based on the prosody and spectrum (frequency) of the voice, rather than the language spoken by the person. This technology uses AI deep learning to perform analysis using 2,548 biomarkers, enabling extremely accurate analysis at the level of medical devices.

●Analysis Modeling Process

Analytical algorithms are created based on labeling by actual doctors (clinical data).

  • Collect voice data
    and doctors’ labeling
    (clinical data)

  • Create
    golden sample

  • Model training
    (Data accumulation
    and features classification)

  • Analytic algorithm

●Used Image

〈Measurable Information〉
  • Under Development

    Cognitive function (MCI), Depression, Fatigue

  • Under Investigation

    Parkinson’s disease, Stress, Anxiety, Energy, etc.

  1. For each disease or detection item, voice data and reference doctors’ labeling (clinical data) need to be collected to develop an analysis model.

Use Cases

This technology can be used in a wide range of fields because of its simple analysis with high accuracy.

  • Dementia/Cognitive Function Decline

    The number of dementia patients in Japan has been increasing with the aging of the population, and has become a major social issue. Early detection is extremely important to prevent progression to dementia.
    Since this technology enables simple analysis of cognitive function decline using only voice, it can be used to monitor changes in cognitive function status through periodic analysis and as a diagnostic support tool at hospitals and health checkup facilities.

  • Stress Check

    Stress checks by companies are problematic because they are conducted once a year, which delays the detection of stress since only the stress condition at the time of the survey can be identified, and because the self-administered questionnaire method makes it difficult to correctly identify the condition if the individual is unaware of it.
    By using this technology, it is possible to identify the stress condition easily and quantitatively. Deterioration of mental health due to teleworking has also become a social issue; this technology can check the stress condition of a speaker during a conference call.

Please feel free to contact us
for technical questions,
collaboration inquiries, etc.

Contact us