Deep Learning-Based Detection of Motor Biomarkers for Autism from Children's Natural Home Video Recordings

Simple item page
dc.authorid0009-0003-8365-1000
dc.contributor.authorFırat, Yelda
dc.contributor.authorKılıçaslan, Yılmaz
dc.contributor.authorSarıkaya, Hasan Ali
dc.contributor.authorYılmaz, Murat Kaan
dc.contributor.authorid41749
dc.date.accessioned2026-06-11T13:35:05Z
dc.date.available2026-06-11T13:35:05Z
dc.date.issued2026-05
dc.departmentFakülteler, Mühendislik, Mimarlık ve Tasarım Fakültesi, Bilgisayar Mühendisliği Bölümü
dc.descriptionScience Citation Index Expanded (SCI-EXPANDED)
dc.description.abstractAutism Spectrum Disorder is a neurodevelopmental disorder with onset in early childhood and its diagnosis often requires clinical processes based on long, subjective observations. Although early diagnosis and intervention can significantly improve developmental outcomes, existing methods are limited in terms of scalability and objectivity. The aim of this study is to develop a hybrid deep learning model that detects Autism Spectrum Disorder with high accuracy by analyzing motor behaviors from videos of children recorded in their natural home environment. In this study, joint coordinates were extracted using the MediaPipe Pose model and spatial, temporal, frequency and coordination-based features were calculated from these data. The features were processed with a hybrid architecture integrating CNN, BiLSTM and attention mechanism. CNN captured spatial patterns, BiLSTM learned the dynamics over time, and the attention mechanism focused on critical movement segments. The model achieves over 97% accuracy on closed datasets and over 83% on public videos such as YouTube and TikTok. These results show that the method performs robustly under both controlled and real-world conditions. The study provides a scalable, objective and clinically applicable screening tool that overcomes the problems of artificial environments and limited data.
dc.identifier.citationFırat, Y., Kılıçaslan, Y., Sarıkaya, H. A. & Yılmaz, M. K. (2026). Deep Learning-based Detection of Motor Biomarkers for Autism from Children's Video Recordings. Journal of Universal Computer Science, 32(4), 519–554. https://doi.org/10.3897/jucs.161202
dc.identifier.doi10.3897/jucs.161202
dc.identifier.endpage554
dc.identifier.issn0948-695X
dc.identifier.issue4
dc.identifier.startpage519
dc.identifier.urihttps://dspace.mudanya.edu.tr/handle/20.500.14362/425
dc.identifier.volume32
dc.identifier.wosWOS:001757035500003
dc.identifier.wosqualityQ3
dc.institutionauthorFırat, Yelda
dc.institutionauthorKılıçaslan, Yılmaz
dc.institutionauthorSarıkaya, Hasan Ali
dc.institutionauthorYılmaz, Murat Kaan
dc.language.isoen
dc.publisherGRAZ UNIV TECHNOLGOY, INST INFORMATION SYSTEMS COMPUTER MEDIA-IICM
dc.relation.journalJournal of Universal Computer Science
dc.relation.publicationcategoryMakale- Uluslararası- Hakemli Dergi- Kurum Öğretim Elemanı
dc.rightsinfo:eu-repo/semantics/openAccess
dc.subjectAutism Spectrum Disorder
dc.subjectMotor Biomarkers
dc.subjectBiLSTM
dc.subjectCNN
dc.subjectAttention
dc.titleDeep Learning-Based Detection of Motor Biomarkers for Autism from Children's Natural Home Video Recordings
dc.typeMakale
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Deep Learning-Based Detection of Motor Biomarkers for Autism from Children's Natural Home Video Recordings.pdf
Size:
1.82 MB
Format:
Adobe Portable Document Format
Description:
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
1.8 KB
Format:
Item-specific license agreed to upon submission
Description:
Download
Collections
Bilgisayar Mühendisliği Bölümü Koleksiyonu