3D Face Reconstruction from 2D Images
First Claim
Patent Images
1. A face reconstruction method, comprising:
- analyzing a plurality of images of a face to find sparse, three-dimensional face features using prior knowledge of a face; and
using said sparse three-dimensional face features to analyze said plurality of images, to find dense three-dimensional features using a data driven approach, without using any prior knowledge.
1 Assignment
0 Petitions
Accused Products
Abstract
A 3D face reconstruction technique using 2D images, such as photographs of a face, is described. Prior face knowledge or a generic face is used to extract spars(e 3D information from the images and to identify image pairs. Bundle adjustment is carried out to determine more accurate 3D camera positions, image pairs are rectified, and dense 3D face information is extracted without using the prior face knowledge. Outliers are removed, e.g., by using tensor voting. A 3D surface is extracted from the dense 3D information and surface detail is extracted from the images.
72 Citations
39 Claims
-
1. A face reconstruction method, comprising:
-
analyzing a plurality of images of a face to find sparse, three-dimensional face features using prior knowledge of a face; and
using said sparse three-dimensional face features to analyze said plurality of images, to find dense three-dimensional features using a data driven approach, without using any prior knowledge. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
-
-
17. A face reconstruction system, comprising:
-
a camera, obtaining a plurality of images of a face; and
a processing part, analyzing said plurality of images, to find sparse, three-dimensional face features using prior knowledge of a face, and using said sparse three-dimensional face features to analyze said plurality of images, to find dense features using a data driven approach, without using any prior knowledge. - View Dependent Claims (18, 19, 20, 21, 22, 23, 24, 25)
-
-
26. A face reconstruction method, comprising:
analyzing a plurality of images from a single camera, to rectify said plurality of images and find three-dimensional information indicative of at least one face from said plurality of images, said analyzing comprising an initial analysis which uses prior knowledge of a face to determine initial features within the images, and a subsequent analysis which uses said initial features of the face to find additional information without using any prior knowledge. - View Dependent Claims (27, 28, 29, 30, 31, 32, 33, 34, 35)
-
36. A method comprising:
-
analyzing a plurality of images of a face to find sparse information about the face;
using the sparse information to find dense information using a data driven approach, said using comprising constraining a scope of the search for dense information using a tensor voting technique.
-
-
37. A face processing method, comprising:
-
analyzing a plurality of images of a face to find matches between said images using prior knowledge of a generic face, said matches used to form sparse information; and
using said matches to form pairs of images;
analyzing said pairs to find a set of dense features using a data driven approach without any prior knowledge of the generic face, said analyzing comprising removing outlier portions from the set of dense features.
-
-
38. A method for automatically reconstructing a 3D face from a plurality of 2D images of a human face, comprising:
-
using prior face knowledge of a generic face in deriving initial camera position estimates;
selecting image pairs and extracting sparse feature points for each of said image pairs;
refining said initial camera position estimates and said sparse feature points;
using a purely data-driven approach in extracting dense 3D point clouds from said image pairs;
merging said dense 3D point clouds into a single 3D cloud;
removing outliers from said single 3D point cloud to form a cleaned 3D point cloud;
fitting a connected surface to the cleaned 3D point cloud, and texture mapping surface detail and color information of a subject'"'"'s face onto said connected surface.
-
-
39. A method for automatically reconstructing a 3D face from a plurality of 2D images of a human face, comprising:
-
using prior face knowledge of a generic face in deriving initial camera position estimates;
selecting image pairs and extracting sparse feature points for each of said image pairs;
refining said initial camera position estimates and said sparse feature points;
using a purely data-driven approach in extracting dense 3D point clouds from said image pairs;
merging said dense 3D point clouds into a single 3D cloud;
using prior face knowledge of a generic face in removing outliers from said single 3D cloud to form a cleaned 3D cloud;
fitting a connected surface to the cleaned 3D point cloud; and
texture mapping surface detail and color information of a subject'"'"'s face onto said surface.
-
Specification