Facial Landmark Detection
Facial landmark detection locates reference points used to describe facial geometry in images or video.
Reference points and benchmark definitions
The 300-W challenge uses the Multi-PIE 68-point landmark markup and a 51-point inner subset. Its error measure is point-to-point distance normalized by inter-ocular distance. That is a benchmark definition, not a universal landmark count for every system. The dataset is for research only, not commercial training.
Landmark geometry is distinct from an interpretation of an expression. Agreement about point locations does not establish what a person felt. Keep a landmark evaluation separate from an Action Unit evaluation or an expression-category evaluation. Each question needs its own defined reference and review procedure.
Open tools and separate terms
OpenFace provides landmark detection, head pose, Action Unit recognition, and eye gaze. Its README directs commercial users to a separate commercial license through CMU. Py-Feat's first version uses 68-point landmarks; its second adds a 478-point MediaPipe mesh. Py-Feat code is MIT-licensed, while pretrained weights have separate terms.
These examples show why counts, tasks, and licensing should be checked separately. They do not describe Mental Edge's implementation or promise identical output. Before comparing tools, specify the visible task and the exact reference annotations being used. Do not treat the presence of more points as evidence of better emotional understanding.
Limits of emotional interpretation
Barrett and colleagues explain that facial configurations vary across people, situations, and cultures. A configuration can express different emotions, or none. Jack and colleagues found differences between Western and East Asian representations of basic expressions. A reply argued that commonalities outweigh differences, so the debate remains active. Neither an Action Unit nor a category score proves feelings, intent, or truthfulness.
The Mental Edge review boundary
Mental Edge surfaces observable expression and movement signals from recorded video for contextual human review. Frame-by-frame scores are associated with seven expression categories, not verified inner states. Use a time-coded signal to revisit the footage and ask an open question. Do not use these outputs for hiring, promotion, diagnosis, deception detection, threat assessment, student scoring, or surveillance.
Review expression signals in context
Mental Edge provides time-coded expression signals from recorded video for human review, not proof of feelings or intent.
Start a free trial