Table 1.

Specifications of the Identified AI Software for Facial Palsy Assessment

Characteristic	Emotrics	Media Pipe Face Mesh	Real-Time Facial Asymmetry Analysis	FaceReader	Affdex
No. of identified studies	30	1	1	4	2
Software validation study	Yes	No	No	Yes	Yes
Platform	Windows	Windows, Linux, IOS	Windows	Windows	Windows, Linux
Software licensing type	Opensource	Opensource	Proprietary	Proprietary	Proprietary
Land markers	68	468	68	468	6
Action unit intensity	No	No	No	Yes	Yes
Landmark measurement	Brow height, marginal reflex distance, commissure excursion, commissure height deviation, smile angle, upper lip deviation, dental show, lower lip deviation	Depends on programming	Oral displacement, eyebrow displacement	No	No
Emotions recognition	No	No	No	Neutral, happy, sad, angry, surprised, scared, disgusted, contempt, arousal, valence	Neutral, happy, sad, angry, surprised, scared, disgusted, contempt, attention, valence, engagement
Unilateral/bilateral	Both	Depends on programming	Unilateral	Both	No study identified
Tutorial	Yes	No	Yes	Yes	Yes
Data storage	Data stored locally on HD	Depends on programming	Data stored locally on HD	Data stored locally on HD	Data stored locally on HD
User interface	Yes	No	Yes	Yes	Yes
Easy to use	Easy	Difficult	Easy	Medium	Medium
URL	https://github.com/dguari1/Emotrics	https://github.com/google/mediapipe	https://www.h.u-tokyo.ac.jp/plastic/contents/RFAA.html	https://www.noldus.com/facereader	https://www.affectiva.com/emotion-ai/ https://imotions.com/products/imotions-lab/modules/fea-facial-expression-analysis/

HD, hard drive.