发明名称 |
Building systems for adaptive tracking of facial features across individuals and groups |
摘要 |
Computer implemented methods for generating a non-transient record of feature locations and/or facial expression parameters characterizing a person's face. A video sequence of a specified individual person is received and a feature locator update model is applied to the video sequence. The feature locator update model is derived by defining a set of training images, generating a set of facial feature displacements for each training image with associated image sample vectors, and training a regularized linear regression which maps from image sample vectors to displacement vectors, wherein the regularization includes a spatial smoothness term within the shape-free sample space. A feature location and/or a facial expression parameter is then extracted, based on the feature update model, characterizing the location, and/or the expression, of a selected set of features of the face of the specified individual person that correspond to an adaptive set of feature locations. |
申请公布号 |
US9104908(B1) |
申请公布日期 |
2015.08.11 |
申请号 |
US201314019748 |
申请日期 |
2013.09.06 |
申请人 |
Image Metrics Limited |
发明人 |
Rogers Michael;Williams Tomos G.;Walker Kevin;Deena Salil |
分类号 |
G06K9/00;G06T13/40 |
主分类号 |
G06K9/00 |
代理机构 |
Sunstein Kann Murphy & Timbers LLP |
代理人 |
Sunstein Kann Murphy & Timbers LLP |
主权项 |
1. A computer implemented method for generating a non-transient record embodying measures of an adaptive set of feature locations characterizing a face of a person, the method comprising:
a. receiving a video sequence that constitutes a physical record of the face of a specified individual person; b. applying a feature locator update model to the video sequence, the feature locator update model derived by steps of:
i. defining a set of training data consisting of training images with associated facial feature locations;ii. generating a set of facial feature displacements for each training image with associated image sample vectors; andiii. training a regularized linear regression which maps from image sample vectors to displacement vectors, wherein the regularized linear regression is characterized by regularization that includes a spatial smoothness term within the shape-free sample space; c. extracting a plurality of locations corresponding to the adaptive set of feature locations; d. based on the feature locator update model, characterizing the locations of features of the face of the specified individual person; and e. storing the locations of features to the non-transient record. |
地址 |
Manchester GB |