Carl S Marshall

US Patent:

6429865, Aug 6, 2002

Filed:

May 6, 1999

Appl. No.:

09/306358

Inventors:

Carl S. Marshall - Portland OR

Assignee:

Intel Corporation - Santa Clara CA

International Classification:

G06T 1700

US Classification:

345420, 345423

Abstract:

Triangle strip length maximization. Embodiments of the invention provide an efficient mechanism to represent triangle meshes as series of one or more triangle strips. In one embodiment, a computer-implemented method first including determining an area of each of a plurality of triangles of a triangle mesh. A distance from the smallest area triangle to every other triangle on the list is determined. An available triangle on the list closest to the smallest area triangle is added to a triangle strip that beings with the smallest area triangle.

US Patent:

6515660, Feb 4, 2003

Filed:

Dec 14, 1999

Appl. No.:

09/459903

Inventors:

Carl S. Marshall - Portland OR
Michael S. MacPherson - Portland OR

Assignee:

Intel Corporation - Santa Clara CA

International Classification:

G06T 1700

US Classification:

345420, 345423, 345428

Abstract:

An apparatus and method is provided for creating a neighbor mesh from a multi-resolution mesh (MRM) representation of a triangulated model. The apparatus and method include computing neighbor data for the neighbor mesh at the highest resolution of the MRM, and creating the neighbor mesh using the computed neighbor data. The apparatus and method further include changing the resolution of the MRM and re-computing the neighbor data at the changed resolution of the MRM, and computing an updated neighbor mesh from the re-computed neighbor data.

US Patent:

6535219, Mar 18, 2003

Filed:

Mar 30, 2000

Appl. No.:

09/539301

Inventors:

Carl S. Marshall - Portland OR
Marc S. Blackstein - Hillsboro OR
Adam T. Lake - Portland OR

Assignee:

Intel Corporation - Santa Clara CA

International Classification:

G06T 1500

US Classification:

345581

Abstract:

A method in a computer system includes identifying faces of a mesh which share an edge and identifying normals of vertices of the faces not included by the edge. The depth components of the normals are multiplied in camera space. When the product of multiplying the depth components of the normals has a first sign, the edge is identified as a silhouette edge.

US Patent:

6552724, Apr 22, 2003

Filed:

May 6, 1999

Appl. No.:

09/306359

Inventors:

Carl S. Marshall - Portland OR 97229

International Classification:

G06T 1700

US Classification:

345420

Abstract:

Triangle strip merging for length maximization of triangle strips representing a triangle mesh. In one embodiment, a computer-implemented method, for each strip, first including determining whether a defining face of the strip has a neighboring face that is a defining face of another strip. A defining face of a particular strip is either a front face, an end face, or an orphan face of the particular strip. Upon determining that a defining face of the strip has a neighboring face that is a defining face of another strip, then the two strips are joined to create a new, larger strip if they are joinable.

US Patent:

6593924, Jul 15, 2003

Filed:

Oct 4, 1999

Appl. No.:

09/410947

Inventors:

Adam T. Lake - Beaverton OR
Carl S. Marshall - Portland OR
Michael S. Rosenzweig - Beaverton OR

Assignee:

Intel Corporation - Santa Clara CA

International Classification:

G06T 1560

US Classification:

345426, 345423

Abstract:

A non-photorealistic image is rendered from an original model in a process which includes selecting a normal in the original model, selecting a vector from a set of vectors, and substituting the vector for the normal to produce a non-photorealistic image corresponding to the original model. The vector may be selected from a subdivision surface that defines a set of vectors, or it may be selected from a set containing a vector pointing substantially toward a virtual light source and a parameterized vector pointing between the virtual light source and a virtual eyepoint.

US Patent:

6593927, Jul 15, 2003

Filed:

Feb 1, 2001

Appl. No.:

09/776129

Inventors:

Daniel P. Horowitz - Issaquah WA
Adam T. Lake - Portland OR
Carl S. Marshall - Portland OR

Assignee:

Intel Corporation - Santa Clara CA

International Classification:

G06T 1500

US Classification:

345473

Abstract:

Rendering a three-dimensional model comprised of three-dimensional data defining a bone and a polygon includes determining a location of a vertex of the polygon relative to the bone, transforming the vertex based on the location to produce a transformed vertex, and rendering the three-dimensional model using the transformed vertex. Determining the location of the vertex includes obtaining a plane that intersects the bone and determining a side of the plane on which the vertex is located.

US Patent:

6608627, Aug 19, 2003

Filed:

Oct 4, 1999

Appl. No.:

09/411680

Inventors:

Carl S. Marshall - Portland OR
Adam T. Lake - Beaverton OR

Assignee:

Intel Corporation - Santa Clara CA

International Classification:

G06T 1500

US Classification:

345619

Abstract:

A two-dimensional image is rendered from data for a three-dimensional model by detecting silhouette features on the three-dimensional model based on the data, and rendering polygons and the silhouette features to produce the two-dimensional image. The silhouette features are detected by detecting edges of polygons that make up the three-dimensional model, storing coordinates of detected polygon edges in an edge list, determining which of the detected polygon edges are silhouette edges, and marking, in the edge list, detected polygon edges that are silhouette edges.

US Patent:

6734853, May 11, 2004

Filed:

Aug 28, 2001

Appl. No.:

09/941501

Inventors:

Oliver A. Heim - Portland OR
Carl S. Marshall - Portland OR

Assignee:

Intel Corporation - Santa Clara CA

International Classification:

G06T 1540

US Classification:

345421, 345623

Abstract:

Processing a scene graph includes applying view frustrum culling to sort models of the scene graph into first and second lists of models, the first list comprising models at least one of within a view frustrum and intersecting with the view frustrum, the second list comprising models outside of the view frustrum, and applying a first selected collision detection process to pairs of models of the first list. The method further includes applying a second selected collision detection process to pairs of models of the second list. In some embodiments, the first collision detection process is more computationally intensive than the second collision detection process.