Brainwave Mood Estimation From Daily Voice Listening
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mood estimation technologies primarily rely on electroencephalograms measured during simple sound stimuli, failing to account for brain responses to daily voice information, which vary significantly in content and impact mental health.
Innovation Solution
A mood estimation program that utilizes machine learning to analyze electroencephalograms from individuals listening to daily voice stimuli, such as news or conversation, generating features like peak latencies and amplitudes to estimate mood scores accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If electroencephalograms are measured during simple sound stimulus tasks, then measurement precision is improved, but adaptability to daily voice information deteriorates
Solution Approach 1:
The patent changes the stimulus parameters from simple standardized sounds to complex daily voice information (news, conversations), while adjusting the machine learning model parameters to handle the increased complexity. This allows the system to maintain measurement precision while adapting to real-world voice variations.
Solution Approach 2:
The estimation model is designed to handle multiple types of voice inputs (news, conversations, various speakers, different acoustic environments) through a universal machine learning framework that processes diverse electroencephalogram responses to different voice stimuli types.
2Adaptability or versatility
If machine learning is used with daily voice information, then adaptability is improved, but device complexity increases
Solution Approach 1:
The patent replaces complex manual mood assessment procedures with an automated machine learning system that processes electroencephalogram data. This substitution of mechanical/manual processes with computational algorithms reduces operational complexity while maintaining high adaptability to different voice inputs.
Solution Approach 2:
The system performs self-adjustment through machine learning, automatically adapting to different voice characteristics and individuals without requiring manual calibration or complex configuration, thereby reducing operational complexity.
3Measurement precision
If electroencephalograms are used for mood estimation, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The patent replaces complex manual mood assessment procedures with automated machine learning analysis of electroencephalogram data. The system automatically processes brain wave responses to voice stimuli and generates mood estimates, eliminating the need for manual psychological evaluation while maintaining high precision.
Solution Approach 2:
The system performs automated mood estimation without requiring operator intervention for data interpretation. The machine learning model self-processes the electroencephalogram features and generates mood scores automatically, simplifying the operational process.
Data Source
AI summary
In the present invention, an estimating device estimates a mood score for a subject by inputting a brain-wave characteristic amount of the subject, when the subject is listening to audio in which a text is read, to an estimation model generated by machine learning which used a plurality of teaching data sets each configured using a combination of a brain-wave characteristic amount of a training test-subject and a mood score of the training test-subject when same was listening to audio in which a text is read.