Please use this identifier to cite or link to this item: http://hdl.handle.net/10125/64148

Research and Design of Autism Smart Diagnosis Information System Based on Chinese Children's Facial Expression Data and Deep Convolution Neural Network

File Size Format  
0328.pdf 1.7 MB Adobe PDF View/Open

Item Summary

Title:Research and Design of Autism Smart Diagnosis Information System Based on Chinese Children's Facial Expression Data and Deep Convolution Neural Network
Authors:Zhao, Wang
Keywords:Data-Driven Smart Health in Asia Pacific
autism
deep convolution neural network
facial expression database
smart diagnostic information system
Date Issued:07 Jan 2020
Abstract:In this paper, the standard facial expression database FER2013 and CK + are used as the main training samples for autism diagnosis model.The facial expression image data of 16 Chinese children were collected as supplementary training samples.We use deep convolution neural network VGG19 and Resnet18 artificial intelligence algorithms to research and develop an smart information system for the diagnosis of autism through facial expression data.Ten normal children and ten autistic children were recruited for the comparative test to verify the accuracy of the system.After testing, the accuracy of facial expression recognition of this system reaches 81.4%.This research is based on the actual business needs of the hospital. The system can diagnose autism as early as possible,and promote the early treatment and rehabilitation of patients, thereby reducing the economic and mental burden of patients. Therefore, this smart information system has good social benefits and application value.
Pages/Duration:10 pages
URI:http://hdl.handle.net/10125/64148
ISBN:978-0-9981331-3-3
DOI:10.24251/HICSS.2020.406
Rights:Attribution-NonCommercial-NoDerivatives 4.0 International
https://creativecommons.org/licenses/by-nc-nd/4.0/
Appears in Collections: Data-Driven Smart Health in Asia Pacific


Please email libraryada-l@lists.hawaii.edu if you need this content in ADA-compliant format.

This item is licensed under a Creative Commons License Creative Commons