Voice-to-Text Abnormality Highlighting for Brain Dysfunction Screening
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies for diagnosing brain dysfunctions, such as dementia, struggle to effectively present the parts of uttered voices that indicate potential abnormalities to users.
Innovation Solution
A program that converts voice input into text, detects abnormal points, and displays them differently on a display unit, allowing users to easily identify these abnormalities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If voice-to-text conversion and abnormality detection are implemented, then detection capability is improved, but user understanding of abnormalities deteriorates due to lack of visual differentiation
Solution Approach 1:
The patent applies local quality by differentiating the display appearance of normal text from abnormal text. Specifically, abnormal character strings are displayed with distinct visual properties (such as different color, background shading, or formatting) compared to normal character strings, allowing users to immediately identify which parts of the converted speech contain abnormalities without losing the contextual information.
2Measurement precision
If detailed text analysis is performed to detect abnormalities, then detection accuracy is improved, but ease of grasping abnormalities deteriorates due to complex presentation
Solution Approach 1:
The patent extracts the abnormal portions from the complete text and presents them separately with visual differentiation. By isolating and highlighting only the abnormal character strings within the full converted speech, the system maintains high detection accuracy while improving ease of grasping, as users can quickly focus on the extracted abnormal parts without being overwhelmed by the entire text.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A program causes a computer to execute processing of: receiving an input of a voice from a subject; converting the input voice into a text; detecting an abnormal point from the text; and displaying the text indicating a character string corresponding to the abnormal point in a display mode different from display modes of other character strings on a display unit when the abnormal point is detected. An input of a voice of a message with respect to a chat group in which a plurality of users including the subject participate is received from the subject, and the message is converted into the text.