Facial recognition and the open mouth problem
First Claim
1. A method of automatic face recognition, comprising providing an electronic representation of a predetermined face having a mouth and being able to adopt open-mouthed or closed mouth states, for electronic face processing to compare with a candidate face, said providing an electronic representation comprising:
- identifying a mouth region within said predetermined face;
applying a topological constraint to said identified mouth region; and
representing said face using said topological constraint to mask said identified mouth region with a mask being substantially invariant to a change in extent of openness of said mouth, thereby to render said representation invariant to a change in an open or closed mouth state.
3 Assignments
0 Petitions
Accused Products
Abstract
A method, for electronic face recognition using geodesic representations, of processing such a geodesic representation of a face having a mouth and being able to adopt open-mouthed or closed mouth states, so that the electronic recognition works regardless of whether the mouth in the representation is open or closed. The method comprises: identifying a mouth region within a face; applying a topological constraint to the mouth region, and representing the face with the constraint. The topological constraint is chosen to render the geodesic representation invariant to the mouth state being open or closed.
-
Citations
33 Claims
-
1. A method of automatic face recognition, comprising providing an electronic representation of a predetermined face having a mouth and being able to adopt open-mouthed or closed mouth states, for electronic face processing to compare with a candidate face, said providing an electronic representation comprising:
-
identifying a mouth region within said predetermined face; applying a topological constraint to said identified mouth region; and representing said face using said topological constraint to mask said identified mouth region with a mask being substantially invariant to a change in extent of openness of said mouth, thereby to render said representation invariant to a change in an open or closed mouth state. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
-
-
9. Apparatus for electronic face recognition by providing an electronic representation of a predetermined face having a mouth and being able to adopt a mouth open state comprising at least an open-mouthed state and a closed mouth state, for comparison with a candidate face for electronic face processing to recognize said candidate face, the apparatus comprising:
-
a mouth region identifier for identifying a mouth region within said predetermined face; and a constraint application unit for applying a topological constraint to said identified mouth region, the apparatus being configured to represent said face and to use said topological constraint to mask said identified mouth region with a mask being substantially invariant to a change in extent of openness of said mouth, thereby to render said representation invariant to a change in a mouth open or closed state. - View Dependent Claims (10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33)
-
Specification