Comparison of marker-less 2D image-based methods for infant pose estimation

In this study we compare the performance of available generic- and specialized infant-pose estimators for a video-based automated general movement assessment (GMA), and the choice of viewing angle for optimal recordings, i.e., conventional diagonal view used in GMA vs. top-down view. We used 4500 annotated video-frames from 75 recordings of infant spontaneous motor functions from 4 to 16 weeks. To determine which pose estimation method and camera angle yield the best pose estimation accuracy on infants in a GMA related setting, the error with respect to human annotations and the percentage of correct key-points (PCK) were computed and compared. The results show that the best performing generic model trained on adults, ViTPose, also performs best on infants. We see no improvement from using specific infant-pose estimators over the generic pose estimators on our infant dataset. However, when retraining a generic model on our data, there is a significant improvement in pose estimation accuracy. This indicates limited generalization capabilities of infant-pose estimators to other infant datasets, meaning that one should be careful when choosing infant pose estimators and using them on infant datasets which they were not trained on. The pose estimation accuracy obtained from the top-down view is significantly better than that obtained from the diagonal view (the standard view for GMA). This suggests that a top-down view should be included in recording setups for automated GMA research.