Video-Audio Neural Network Ensemble For Comprehensive Screening Of Autism Spectrum Disorder in Young Children DOI Open Access
Shreyasvi Natraj, Nada Kojovic, Thomas Maillart

et al.

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: July 3, 2023

ABSTRACT A timely diagnosis of autism is paramount to allow early therapeutic intervention in preschoolers. Deep Learning (DL) tools have been increasingly used identify specific autistic symptoms, and offer promises for automated detection at an age. Here, we leverage a multi-modal approach by combining two neural networks trained on video audio features semi-standardized social interactions sample 160 children aged 1 5 years old. Our ensemble model performs with accuracy 82.5% (F1 score: 0.816, Precision: 0.775, Recall: 0.861) ASD screening. Additional combinations our were developed achieve higher specificity (92.5%, i.e., few false negatives) or sensitivity (90%, i.e. positives). Finally, found relationship between the network modalities versus characteristics, bringing evidence that implementation was effective taking into account different are currently standardized under gold standard assessment.

Language: Английский

Predicting early ASD traits of adults and toddlers using machine learning and deep learning with explainable AI and optimization DOI
Md. Ashiqur Rahman,

Md. Mamun Hossain,

Sondip Poul Singh

et al.

Neural Computing and Applications, Journal Year: 2025, Volume and Issue: unknown

Published: March 27, 2025

Language: Английский

Citations

0
Autism Spectrum Disorder Detection in Children Via Deep Learning Models Based on Facial Images DOI Creative Commons

Bawer Khan,

Arslan Akram, Sohail Masood Bhatti

et al.

Bulletin of Business and Economics (BBE), Journal Year: 2024, Volume and Issue: 13(1)

Published: March 25, 2024

Autism spectrum disorder (ASD) is a complicated psychiatric disease that causes difficulty in communicating with others, and restricted behavior, speech, as well nonverbal interaction. Children autism have unique facial characteristics distinguish them from ordinarily developing children. Therefore, there requirement for precise automated system capable of early detection children, yielding accurate results. The objective this research to assist both families psychiatrists diagnosing through straightforward approach. Specifically, the study employs deep learning method utilizes experimentally validated features. technique involves convolutional neural network along transfer autism. MobileNetv2, Xception, ResNet-50, VGG16 DenseNet-121 were pretrained models used evaluation these utilized dataset sourced Kaggle, comprising 2,940 images. We evaluated five using standard measures like recall, precision, accuracy, F1 score, ROC curve. proposed model outperformed existing models, 96% accuracy rate. With respect performance evaluation, exhibited superiority over most recent models. Our possesses capability support healthcare professionals validating precision their initial screening Spectrum Disorders (ASDs) pediatric patients.

Language: Английский

Citations

3
Assessing Autistic Traits in Toddlers Using a Data-Driven Approach with DSM-5 Mapping DOI Creative Commons
Neda Abdelhamid,

Rajdeep Thind,

Heba Mohammad

et al.

Bioengineering, Journal Year: 2023, Volume and Issue: 10(10), P. 1131 - 1131

Published: Sept. 27, 2023

Autistic spectrum disorder (ASD) is a neurodevelopmental condition that characterises range of people, from individuals who are not able to speak others have good verbal communications. The affects the way people see, think, and behave, including their communications social interactions. Identifying autistic traits, preferably in early stages, fundamental for clinicians expediting referrals, hence enabling patients access required healthcare services. This article investigates various ASD behavioral features toddlers proposes data process using machine-learning techniques. aims this study were identify can help detect map these neurodevelopment areas Diagnostic Statistical Manual Mental Disorders (DSM-5). To achieve aims, proposed assesses several feature selection techniques, then constructs classification model based on chosen features. empirical results show during screening toddlers, cognitive related communications, interactions, repetitive behaviors most relevant ASD. For algorithms, predictive accuracy Bayesian network (Bayes Net) logistic regression (LR) models derived subsets consistent pinpointing suitability ML techniques predicting

Language: Английский

Citations

4
Developing an Artificial Intelligence Based Model for Autism Spectrum Disorder Detection in Children DOI Creative Commons
Chitta Hrudaya Neeharika,

Yeklur Mohammed Riyazuddin

Journal of Advanced Research in Applied Sciences and Engineering Technology, Journal Year: 2023, Volume and Issue: 32(1), P. 57 - 72

Published: Aug. 19, 2023

Sensory difficulties, such as an over or under responsiveness to noises, smells, touch, are frequently present in individuals with Autism Spectrum Disease (ASD), a neurodevelopmental disorder. The condition's primary cause is hereditary, however early diagnosis and therapy can assist. Traditional clinical procedures may be expensive time consuming, but current history, deep learning based sophisticated has emerged supplement them. goal of this study streamline the diagnostic procedure by identifying most important characteristics automating them using existing classification methods. We have looked at datasets including toddlers, kids, teens, adults autism spectrum To find highest performing feature set for these four ASD datasets, we compared state-of-the-art categorization selection Across adults, our experiments reveal that multilayer perceptron (MLP) classifier achieves 100% accuracy fewest possible features. also determine proposed approach ranks across all datasets.

Language: Английский

Citations

4
Enhanced Special Needs Assessment: A Multimodal Approach for Autism Prediction DOI Creative Commons
Suseela Sellamuthu,

Sharon Rose

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 121688 - 121699

Published: Jan. 1, 2024

Language: Английский

Citations

1
High-Performance Grape Disease Detection Method Using Multimodal Data and Parallel Activation Functions DOI Creative Commons
Ruiheng Li, Jiarui Liu,

Binqin Shi

et al.

Plants, Journal Year: 2024, Volume and Issue: 13(19), P. 2720 - 2720

Published: Sept. 28, 2024

This paper introduces a novel deep learning model for grape disease detection that integrates multimodal data and parallel heterogeneous activation functions, significantly enhancing accuracy robustness. Through experiments, the demonstrated excellent performance in detection, achieving an of 91%, precision 93%, recall 90%, mean average (mAP) 56 frames per second (FPS), outperforming traditional models such as YOLOv3, YOLOv5, DEtection TRansformer (DETR), TinySegformer, Tranvolution-GAN. To meet demands rapid on-site this study also developed lightweight mobile devices, successfully deployed on iPhone 15. Techniques structural pruning, quantization, depthwise separable convolution were used to reduce model’s computational complexity resource consumption, ensuring efficient operation real-time performance. These achievements not only advance development smart agricultural technologies but provide new technical solutions practical tools detection.

Language: Английский

Citations

1
Efficient machine learning models across multiple datasets for autism spectrum disorder diagnoses DOI
Amr E. Eldin Rashed, Waleed M. Bahgat, Ali Mohammed Saleh Ahmed

et al.

Biomedical Signal Processing and Control, Journal Year: 2024, Volume and Issue: 100, P. 106949 - 106949

Published: Oct. 4, 2024

Language: Английский

Citations

1
An Effective Model Of Autism Spectrum Disorder Using Machine Learning DOI Open Access
Razieh Asgarnezhad,

Karrar Ali Mohsin Alhameedawi,

Hani Akram Mahfoud

et al.

Indonesian Journal of Electrical Engineering and Informatics (IJEEI), Journal Year: 2023, Volume and Issue: 11(2)

Published: April 15, 2023

Autism spectrum disorder (ASD) is one of the most common diseases that affect human nerves and cause a decrease in intelligence comprehension person. This disease group various disorders are characterized by poor social behavior communication. It affects all age groups, including adults, adolescents, children, elderly, but symptoms this always appear their early years. ASD suffer from problems, important which data loss, low quality, extreme values. makes process diagnosing early. Our goals research to solve problems. The cussent authors proposed technical model solves We used ensemble techniques include Bayesian Boosting, Classification Regression, Polynomial Binominal Classification. also classification CHAID, Decision Stump, Tree (Weight-Based), Gradient Boosted Trees, ID3. proven has obtained highest search accuracy reached 100% as well we have f1 measurement 100%. proves our work superior its peers.

Language: Английский

Citations

1
Medication Usage Record-Based Predictive Modeling of Neurodevelopmental Abnormality in Infants under One Year: A Prospective Birth Cohort Study DOI Open Access
Tianyi Zhou, Yaojia Shen, Jinlang Lyu

et al.

Healthcare, Journal Year: 2024, Volume and Issue: 12(7), P. 713 - 713

Published: March 24, 2024

Early identification of children with neurodevelopmental abnormality is a major challenge, which crucial for improving symptoms and preventing further decline in abnormality. This study focuses on developing predictive model maternal sociodemographic, behavioral, medication-usage information during pregnancy to identify infants abnormal neurodevelopment before the age one. In addition, an interpretable machine-learning approach was utilized assess importance variables model. this study, artificial neural network models were developed five areas first year life achieved good efficacy fine motor problem solving, median AUC = 0.670 (IQR: 0.594, 0.764) 0.643 0.550, 0.731), respectively. The final abnormalities any energy region one-year-old also prediction performance. sensitivity 0.700 0.597, 0.797), 0.821 0.716, 0.833), accuracy 0.721 0.696, 0.739), specificity 0.742 0.680, 0.748). methods suggest that exposure drugs such as acetaminophen, ferrous succinate, midazolam affects development specific offspring life. established under one underscored value medication outcomes offspring.

Language: Английский

Citations

0
Computational Interpersonal Communication Model for Screening Autistic Toddlers: A Case Study of Response-to-Name DOI
Wei Nie, Bingrui Zhou, Zhiyong Wang

et al.

IEEE Journal of Biomedical and Health Informatics, Journal Year: 2024, Volume and Issue: 28(6), P. 3683 - 3694

Published: April 16, 2024

Interpersonal communication facilitates symptom measures of autistic sociability to enhance clinical decision-making in identifying children with autism spectrum disorder (ASD). Traditional methods are carried out by practitioners assessment scales, which subjective quantify. Recent studies employ engineering technologies analyze children's behaviors quantitative indicators, but these only generate specific rule-driven indicators that not adaptable diverse interaction scenarios. To tackle this issue, we propose a Computational Communication Model (CICM) based on psychological theory represent dyadic interpersonal as stochastic process, providing scenario-independent theoretical framework for evaluating sociability. We apply CICM the response-to-name (RTN) 48 subjects, including 30 toddlers ASD and 18 typically developing (TD), design joint state transition matrix indicators. Paired machine learning, our proposed CICM-driven achieve consistencies 98.44% 83.33% RTN expert ratings diagnosis, respectively. Beyond outstanding screening results, also reveal interpretability between statistical analysis.

Language: Английский

Citations

0