Using DeepLabCut for tracking body landmarks in videos of children with dyskinetic cerebral palsy: a working methodology

Abstract

Background: Markerless motion tracking is a promising technique to capture human movements and postures. It could be a clinically feasible tool to objectively assess movement disorders within dyskinetic cerebral palsy (CP). Here, we aim to evaluate markerless motion tracking on clinically recorded video data , i.e. to assess its accuracy against manual labeling. Method: 94 video recordings of 33 participants (dyskinetic CP, 8-23 years; GMFCS IV-V) from a previous clinical trial were used. Twenty-second clips were cut during lying down. Video image resolution was 0.4 cm per pixel. Automatic tracking was performed in DeepLabCut. We evaluated a model that was pretrained on a human healthy adult data set with an increasing number of manually labeled frames (0, 1, 2, 6, 10, 15 and 20 frames per video). To assess generalizability, we used 80% of videos for the model development and evaluated the generalizability of the model using the remaining 20%. For evaluation the mean absolute error (MAE) between DeepLabCuts prediction of the position of body points and manual labels was calculated. Results: Using just the pre-trained adult human model yielded a MAE of 121 pixels. An MAE of 4.5 pixels (about 1.5 cm) could be achieved by adding 15-20 manual labels. When applied to unseen video clips (i.e. generalization set), the MAE was 33 pixels with a dedicated model trained on 20 frames per videos. Conclusion: Markerless motion tracking with a standard pre-trained model is insufficiently accurate to automatically assess movement disorders in dyskinetic CP. However, manually adding labels improves the model performance substantially. In addition, the methodology proposed within our study is applicable to check the accuracy of DeepLabCut application within other clinical data sets.

Automatic extraction of upper-limb kinematic activity using deep learning-based markerless tracking during deep brain stimulation implantation for Parkinson’s disease: A proof of concept study

“…Our contributions could elaborate on this use by categorizing aspects of gait based on Euclidean distance movement epochs, thereby presenting an entirely objective option that performs well even in the absence of strictly-controlled room conditions required for videotaped observational gait analysis [ 25 ]. In pediatric patients, rehabilitation scientists have used a similar methodology to evaluate dyskinetic cerebral palsy symptoms, which could also be expanded upon by our findings [ 61 ].…”