Research Institute of Visual Computing

Select a project of interest using the advanced search:

Facial Movement in 3 Dimensions: Average Templates of Lip Movement in Adults

Authors: Popat H, Richmond S, Marshall D, Rosin PL

Abstract:

To construct 3-dimensional (3D) templates of average lip movement based on a group of normal healthy subjects.

SUBJECTS AND METHODS:

One hundred fifteen white subjects were asked to perform 2 reproducible verbal gestures (/puppy/ and /rope/) in a normal relaxed manner. The sequences were captured using a noninvasive, 3D motion scanner (3dMDFace Dynamic System). Mesh-registration software was used to align sequential facial shells to a standardized reference plane. Lip movement during the verbal facial gestures was quantified using the displacement vectors (x, y, z coordinates) of 6 lip landmarks of sequential 3D facial shells subtracted from the reference or rest shell. Descriptive statistics were used to build references for average lip movement, and independent t tests were used to compare lip movement between men and women.

RESULTS:

Mean reference movements were created for lip opening, lip stretch, and lip purse that corresponded to the visemes /pu/ /ppy/ and /rope/, respectively. There were statistically significant differences in the lip movement between genders for the visemes /pu/ and /ppy/, although when quantified these were found not to be of clinical significance. Men favored left-sided movement and women right-sided movement, although only slightly--again, when quantified, this was considered to be within symmetrical limits.

CONCLUSION:

It was possible to quantify and create normal templates of lip movement for the words /puppy/ and /rope/. Men and women show similar standardized lip movements for these 2 words.

Link to Paper

Authors

Prof. Paul Rosin

Prof. Paul Rosin

Various aspects of computer vision, including 2D and 3D facial analysis and synthesis.

Prof. David Marshall

Prof. David Marshall

Computer Vision including Reverse Engineering, Automated Inspection, Dynamic 2D/3D Facial Analysis.