Personalized ASD Diagnosis Mapping Model Using Global AI
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current ASD diagnosis systems rely heavily on continuous observation by medical experts and are hindered by the reluctance of children and parents to visit hospitals, long waiting times, and the inability of AI models to reflect individual patient patterns and medical expert experiences.
Innovation Solution
A personalization-based telemedicine diagnosis method and apparatus that uses social-interaction-inducing content to assist medical experts in ASD diagnosis, generating a personalized ASD diagnosis mapping model using a pretrained global ASD diagnosis model and limited diagnostic values from medical experts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a pretrained global ASD diagnosis model is used, then diagnostic coverage and generalization are improved, but individual patient characteristics and medical expert experiences are not reflected
Solution Approach 1:
The diagnosis system is segmented into two hierarchical levels: a global pretrained model that provides general diagnostic capabilities, and a personalized mapping model that adapts to individual patients and medical experts. This segmentation allows the system to maintain broad diagnostic coverage while simultaneously achieving high individual diagnosis accuracy through the specialized mapping model.
Solution Approach 2:
The system merges the pretrained global ASD diagnosis model with a personalized mapping model into an integrated diagnostic framework. The global model processes input images to generate preliminary diagnostic values, which are then transformed by the mapping model to reflect individual patient characteristics and medical expert experiences, combining the strengths of both approaches.
2Measurement precision
If continuous observation by medical experts is used, then diagnostic accuracy is improved, but time consumption and hospital visit requirements increase
Solution Approach 1:
The system enables automated self-service diagnosis where the AI model independently analyzes input images and generates diagnostic results without requiring continuous medical expert observation. The personalized mapping model allows the system to adapt to medical expert preferences automatically, reducing the need for manual intervention while maintaining high diagnostic accuracy.
Solution Approach 2:
The system replaces the mechanical process of continuous human observation with an automated AI-based image analysis system. The pretrained global model and personalized mapping model together perform diagnostic functions that previously required sustained medical expert attention, significantly reducing diagnosis time while maintaining or improving accuracy.
3Productivity
If AI models are used to assist diagnosis, then productivity is improved, but susceptibility of medical experts to AI results is reduced due to lack of personalization
Solution Approach 1:
The mapping model is designed to be dynamic and adaptable, adjusting its parameters based on individual medical expert preferences and experiences. This dynamic personalization increases medical expert susceptibility to AI results by aligning the system's output with each expert's diagnostic style and expectations, thereby improving both productivity and reliability.
Solution Approach 2:
The system incorporates feedback mechanisms where medical expert interactions with diagnostic results are used to refine and personalize the mapping model. This feedback loop enhances the model's ability to reflect individual expert experiences, increasing trust and acceptance while maintaining high diagnostic efficiency.
Data Source
AI summary
Disclosed herein is a method for assisting in Autism Spectrum Disorder (ASD) diagnosis. The method includes transmitting social-interaction-inducing content, receiving input images containing a response of an assessment subject to the social-interaction-inducing content, receiving an ASD diagnosis result for a preset number of input images, among the received input images, and outputting a diagnostic assistive result for the received input images using ASD diagnosis input for the preset number of input images and a pretrained global ASD diagnosis model.