Facial Feature Analysis for Non-Invasive Dementia Determination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for dementia diagnosis, such as amyloid PET and cerebrospinal fluid biomarkers, are invasive or expensive, and existing AI-based methods lack sufficient accuracy.
Innovation Solution
A disease relevance determination apparatus and program that uses both normal and smiling face data to determine dementia through feature point analysis, utilizing deep learning to compare weighted feature points between reference and test subjects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If amyloid PET is used for dementia diagnosis, then determination accuracy is improved, but cost increases and radiation exposure occurs
Solution Approach 1:
The patent uses facial image data as a non-invasive copy or alternative to invasive biomarkers and PET scans. By analyzing facial features through deep learning, the system replicates the diagnostic capability of expensive invasive tests without requiring physical sampling or radiation exposure, thus resolving the contradiction between accuracy and harmful factors
Solution Approach 2:
The patent employs standard digital camera equipment and publicly available facial images instead of expensive medical imaging equipment. This substitution with inexpensive, readily available resources achieves diagnostic accuracy without the high costs and harmful effects associated with traditional methods
2Measurement precision
If cerebrospinal fluid biomarkers are used for dementia diagnosis, then determination accuracy is improved, but invasiveness increases due to needle insertion
Solution Approach 1:
The patent replaces invasive cerebrospinal fluid sampling with non-invasive facial image analysis. The system creates a digital representation of neurological status through facial feature patterns, eliminating the need for needle insertion while maintaining diagnostic accuracy through deep learning algorithms
Solution Approach 2:
The patent substitutes the mechanical invasive procedure of lumbar puncture with a computational optical system. Instead of physically inserting needles to collect fluid samples, the system uses cameras to capture facial images and computational algorithms to analyze them, replacing mechanical invasion with optical and information processing methods
3Measurement precision
If traditional dementia diagnosis methods are used, then determination accuracy is improved, but cost increases
Solution Approach 1:
The patent uses inexpensive digital cameras and standard computational resources instead of expensive medical imaging equipment. By leveraging freely available facial images and open-source deep learning frameworks, the system achieves diagnostic accuracy at a fraction of the cost of traditional methods
Solution Approach 2:
The patent replicates the diagnostic function of expensive medical tests using inexpensive facial image analysis. The system creates a low-cost alternative that copies the essential diagnostic capability of PET scans and biomarker analysis through analysis of facial feature patterns associated with neurological conditions
Data Source
AI summary
[Problem] To provide: a dementia determination device which enables the accurate determination of dementia in a non-invasive, inexpensive and simple manner; and a dementia determination program.[Solution] Provided is a disease relevance determination program for making a computer execute the below-mentioned steps to determine the disease relevance in a test subject, in which the steps include: a feature point acquisition step for acquiring weighed feature point data X in a reference subject using default face photo data of a smiley face of the reference subject which have been acquired in advance; a distinction data acquisition step for acquiring data A correspo DISTINCTION he weighed feature point from the face photo data for DISTINCTION ject; and a determination step for comparing the data A with the data X to determine the disease relevance in the test subject. Also provided is a disease relevance determination program which is provided with a computer in which the disease relevance determination program is loaded.