Skip to main content
. 2020 Sep 30;10:333. doi: 10.1038/s41398-020-01015-w

Table 5.

Stereotyped behaviours.

Reference Focus N participants Age Input data/device used Method used Dataset
Hashemi et al.95 Behaviour analysis 6 ASD, 14 TD 16–30 months Video, iPad IntraFace Own dataset
Hashemi et al.92 Sharing interest, visual tracking, and disengagement of attention detection 3 ASD, 3 TD 6–15 months Video, two GoPro HD cameras Histogram of Orientated Gradients (HOG) and SVM Own dataset
Hashemi et al.93 Behaviour analysis 12 ASD 5–16 months GoPro Hero HD HOG and SVM Own dataset
Bidwell et al.94 Behaviour analysis 15–30 months Video, Camera and Microsoft Kinect Omron OKAO Vision Library Multimodal Dyadic Behaviour (MMDB) Dataset
Campbell et al.96 Atypical orienting and attention behaviours for behavioural observation 22 ASD, 82 TD or DD 16–31 months Tablet Device IntraFace Own dataset
Hashemi et al.97 Engagement, name-call responses, and emotional responses 15 ASD, 18 TD 16–31 months Video, iPad IntraFace Own dataset
Wang et al.98 Attention monitoring for diagnosis 5 ASD, 12 TD Video, two RGB cameras Microsoft SDK Own dataset
Bovery et al.99 Attention monitoring for behavioural assessment 22 ASD, 82 TD 16–31 months Video, iPad IntraFace Own dataset
Rajagopalan and Goecke100 Self-stimulatory behaviour detection YouTube videos Histogram of Dominant Motions (HDM) Self-stimulatory Behaviour Dataset, UCF101 and Weizmann Datasets
Rajagopalan et al.101 Self-stimulatory behaviour detection YouTube videos Space Time Interest Points (STIP) with Harris3D detectors in a BOW framework Self-stimulatory Behaviour Dataset
Rajagopalan102 Self-stimulatory behaviour detection YouTube videos Motion trajectories Self-stimulatory Behaviour Dataset, UCF101, Hollywood 2 Datasets
Winoto et al.103 Behaviour analysis 4 ASD, 4 TD Microsoft Kinect v2
Feil-Seifer and Matarić104 Interaction with robots for behaviour analysis 8 ASD Video, camera Heuristics Own dataset
Moghadas and Moradi105 Interaction with robots for diagnosis 8 ASD, 8 TD

ASD: 2.1–4.1 years

TD: 2.11–7.6 years

 Video, RobotParrot and two cameras Kernelised Correlation Filter (KFC) and cosine similarity and SVM Own dataset