Automated mapping of facial images to animation wireframes topologies

Claim 2

A computerized method for mapping a digitized target image of a face to a wireframe topology stored in a memory in a computer system, where the image is composed of a plurality of pixels arranged in a regular pattern, and the wireframe topology is composed of a plurality of nodes connected by lines, comprising the steps of: detecting a plurality of facial features contained in both the image and wireframe topology; computing a bounding box from said plurality of detected facial features; cropping a reference facial image using the bounding box, wherein the reference facial image is a composite image of a plurality of facial images; cropping the digitized target facial image using the bounding box; registering the cropped reference facial image with the cropped target facial image to determine a displacement field using a spline-based image registration; and applying the displacement field to a reference wireframe topology to map the target facial image to the wireframe topology.

Claim 3

The method of claim 1 wherein the reference facial image is a composite image of a plurality of facial images.

Claim 4

A computerized method for mapping a digitized target image of a face to a wireframe topology stored in a memory in a computer system, where the image is composed of a plurality of pixels arranged in a regular pattern, and the wireframe topology is composed of a plurality of nodes connected by lines, comprising the steps of: detecting a plurality of facial features contained in both the image and wireframe topology by the steps of: determining a plurality of facial landmark nodes in the wireframe topology corresponding to the detected facial features; computing a transform between the facial features and the landmark nodes; applying the transform to the wireframe topology to generate a reference wireframe topology for the reference facial image; computing a bounding box from said plurality of detected facial features; cropping a reference facial image using the bounding box: cropping the digitized target facial image using the bounding box; registering the cropped reference facial image with the cropped target facial image to determine a displacement field; and applying, the displacement field to a reference wireframe topology to map the target facial image to the wireframe topology.

Claim 5

A computerized method for mapping a digitized target image of a face to a wireframe topology stored in a memory in a computer system, where the image is composed of a plurality of pixels arranged in a regular pattern, and the wireframe topology is composed of a plurality of nodes connected by lines, comprising the steps of: detecting a plurality of facial features contained in both the image and wireframe topology by the steps of: determining a plurality of facial landmark nodes in the wireframe topology corresponding to the detected facial features; computing a transform between the facial features and the landmark nodes; applying the transform to the wireframe topology to generate a reference wireframe topology for the reference facial image; computing a bounding box from said plurality of detected facial features; cropping a reference facial image using the bounding box: cropping the digitized target facial image using the bounding box; registering the cropped reference facial image with the cropped target facial image to determine a displacement field; and applying, the displacement field to a reference wireframe topology to map the target facial image to the wireframe topology.

Claim 6

The method of claim 3, wherein the transform is a global affine transform with shifting and scaling of the pixels in x- and y- directions.

Claim 7

The method of claim 1 wherein the displacement field is expressed as the displacement of spline control vertices between the cropped facial images.

Claim 8

A computerized method for mapping a digitized image of a face to a wireframe topology stored in a memory in a computer system, where the image is composed of a plurality of pixels arranged in a regular pattern, and the wireframe topology is composed of a plurality of nodes connected by lines, comprising the steps of: detecting a plurality of facial features from the plurality of pixels of a reference facial image, where the reference facial image is a composite image of a plurality of facial images, and the target facial image is registered with the composite reference facial image; determining a plurality of facial landmark nodes in the wireframe topology corresponding to the detected facial features; computing a transform between the facial features and the landmark nodes; applying the transform to the wireframe topology to generate a reference wireframe topology for the reference facial image; cropping the reference facial image using a bounding box; cropping a target facial image using the bounding box; registering the cropped reference facial image with the cropped target facial image to determine a displacement field; and applying the displacement field to the reference wireframe topology to map the target facial image to the wireframe topology.

Claim 9

A computerized method for mapping a digitized image of a face to a wireframe topology stored in a memory in a computer system, where the image is composed of a plurality of pixels arranged in a regular pattern, and the wireframe topology is composed of a plurality of nodes connected by lines, comprising the steps of: detecting a plurality of facial features from the plurality of pixels of a reference facial image, where the reference facial image is a composite image of a plurality of facial images, and the target facial image is registered with the composite reference facial image; determining a plurality of facial landmark nodes in the wireframe topology corresponding to the detected facial features; computing a transform between the facial features and the landmark nodes; applying the transform to the wireframe topology to generate a reference wireframe topology for the reference facial image; cropping the reference facial image using a bounding box; cropping a target facial image using the bounding box; registering the cropped reference facial image with the cropped target facial image to determine a displacement field; and applying the displacement field to the reference wireframe topology to map the target facial image to the wireframe topology.