Abstract:General movement (GM) assessment (GMA) is an internationally recognised tool for the early screening and diagnosis of neurodevelopmental abnormalities in high-risk infants. Traditional GMA requires multiple internationally certified doctors, which is subjective and time-consuming and therefore limits its widespread use, especially among neonates. Quantifying and accelerating GMA can reduce artificial and experience-related interference and provide more accurate judgments. Using transfer learning from MediaPipe BlazePose and a homemade three-view high-definition video tracking system, we built a safe video recording environment for neonates; 103 sets of three-view GM videos of 96 neonates were captured using 23-key point analysis and computer vision machine learning technology. These were used to accurately identify and track fine movements in the neonate twisting motion stage (40 weeks of corrected gestational age) to classify different neonatal GM types. In this work, for the front-view video data (including 37 normal GMs and 37 abnormal GMs randomly selected from 66 abnormal GMs), after extracting 17 time-domain and 10 frequency-domain features from the time series of key point spatial coordinates and pre-set joint angles, eight supervised classification algorithms-i. e. decision trees, naive Bayes, support vector machine, logistic regression, k-nearest neighbour, random forest, gradient boosting, and voting classifier-were used to differentiate abnormal from normal GMs. Logistic regression demonstrated the best performance; the highest prediction accuracy (93.33 %) indicated a consistent agreement rate of 93.33 % relative to the physician diagnosis; the precision (93.75), recall (93.75), F1-score (93.33 %), and the area under the receiver operating characteristic curve (1.000) gave further validation. Thus, video tracking with unlabelled visual sensors and automatic quantification with artificial intelligence make GMA more objective and reliable, thereby broadening its use.

RGB-D Videos-Based Early Prediction of Infant Cerebral Palsy Via General Movements Complexity

Automatically Measure the Quality of Infants’ Spontaneous Movement Via Videos to Predict the Risk of Cerebral Palsy

Automating General Movements Assessment with Quantitative Deep Learning to Facilitate Early Screening of Cerebral Palsy

A pose estimation for motion tracking of infants cerebral palsy

Assessing infant risk of cerebral palsy with video-based motion tracking

A Training-Free Infant Spontaneous Movement Assessment Method for Cerebral Palsy Prediction Based on Videos.

Preterm Infant General Movements Assessment Via Representation Learning

Three-Dimensional Pose Estimation of Infants Lying Supine Using Data from a Kinect Sensor with Low Training Cost

Automatic Quantitative Intelligent Assessment of Neonatal General Movements with Video Tracking

Cost-effective Video-based Poor Repertoire Detection for Preterm Infant General Movement Analysis.

Towards Reliable, Automated General Movement Assessment for Perinatal Stroke Screening in Infants Using Wearable Accelerometers

A Spatio-temporal Attention-based Model for Infant Movement Assessment from Videos

CP-AGCN: Pytorch-based Attention Informed Graph Convolutional Network for Identifying Infants at Risk of Cerebral Palsy

RGB-D Camera based Human Limb Movement Recognition and Tracking in Supine Positions.

Automated detection of abnormal general movements from pressure and positional information in hospitalized infants

Weakly Supervised Online Action Detection for Infant General Movements

Evaluating Explainable AI Methods in Deep Learning Models for Early Detection of Cerebral Palsy

Machine learning approaches to evaluate infants' general movements in the writhing stage—a pilot study

Automatic Classification of General Movements in Newborns

Cerebral Palsy Prediction with Frequency Attention Informed Graph Convolutional Networks

Deep-learning for automated markerless tracking of infants general movements