The future of General Movement Assessment: The role of computer vision and machine learning - A scoping review.
Nelson Silva, Dajie Zhang, Tomas Kulvicius and 10 others
PMID 33571849WHAT IT FOUND
Automated video analysis of infant general movements is advancing, but no current tool matches trained human assessment.
It may support early referral later, not replace clinical evaluation.
Key findings
01No current automated video-based general movement assessment tool yet matches the performance of human experts.
02The reviewed studies use different setups, tracking methods, and datasets, so their performance cannot be compared across studies.
03Automated general movement tools are intended to complement human assessment, not replace it.
STILL TO COME
How it was doneWhat they foundWhat it means for PTs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The review only included English publications from 2010 onwards, so relevant work in other languages or older papers may be missing. The 40 studies used different cameras, movement tracking methods, features, and datasets, so their accuracy cannot be compared directly. Many studies did not report key infant characteristics such as gestational age, and some included infants outside the age range for general movements. Most studies did not report whether the movement labels were made by certified general movement assessment experts or whether those labels were reliable. No large public dataset with expert general movement labels was available for independent testing. The review maps technological approaches; it does not test whether automated general movement assessment improves diagnosis, referral, or patient outcomes.
Declared interests
The authors declare no conflicts of interest.
The easy way to misread this
Do not read the growth of automated video analysis as evidence that it can replace human general movement assessment. The review found no current automated solution defeats human experts, and these tools are intended to complement human assessment.