17919460. Systems and Methods for Compression of Three-Dimensional Volumetric Representations simplified abstract (Google LLC)

Systems and Methods for Compression of Three-Dimensional Volumetric Representations

Organization Name

Google LLC

Inventor(s)

Danhang Tang of San Francisco CA (US)

Saurabh Singh of Mountain View CA (US)

Cem Keskin of San Francisco CA (US)

Phillip Andrew Chou of Bellevue WA (US)

Christian Haene of Berkeley CA (US)

Mingsong Dou of Cupertino CA (US)

Sean Ryan Francesco Fanello of San Francisco CA (US)

Jonathan Taylor of San Francisco CA (US)

Andrea Tagliasacchi of Toronto (CA)

Philip Lindsley Davidson of Arlington MA (US)

Yinda Zhang of Dale City CA (US)

Onur Gonen Guleryuz of San Francisco CA (US)

Shahram Izadi of Tiburon CA (US)

Sofien Bouaziz of Los Gatos CA (US)

Systems and Methods for Compression of Three-Dimensional Volumetric Representations - A simplified explanation of the abstract

This abstract first appeared for US patent application 17919460 titled 'Systems and Methods for Compression of Three-Dimensional Volumetric Representations

Simplified Explanation

The patent application describes a system and method for encoding and decoding the textures and geometry of a three-dimensional volumetric representation.

• The system can obtain voxel blocks from a three-dimensional representation of an object.
• These voxel blocks are then encoded using a machine-learned voxel encoding model to obtain encoded voxel blocks.
• The encoded voxel blocks are decoded using a machine-learned voxel decoding model to obtain reconstructed voxel blocks.
• Based on these reconstructed voxel blocks, a reconstructed mesh representation of the object is generated.
• Textures associated with the voxel blocks are encoded according to an encoding scheme and based on the reconstructed mesh representation to obtain encoded textures.

Potential Applications

This technology has various potential applications, including:

• 3D modeling and animation: The system can be used to efficiently encode and decode the textures and geometry of three-dimensional objects, making it useful in fields such as gaming, virtual reality, and computer-generated imagery.
• Medical imaging: The system can be applied to encode and decode volumetric representations of medical scans, enabling efficient storage and transmission of medical imaging data.
• Architectural design: The technology can be utilized to encode and decode three-dimensional models of buildings and structures, facilitating efficient visualization and collaboration in architectural design processes.

Problems Solved

The technology addresses several problems in the field of three-dimensional representation encoding and decoding, including:

• Efficient encoding: The system provides a method for encoding voxel blocks and textures in a manner that reduces storage and transmission requirements.
• Accurate decoding: The machine-learned voxel decoding model ensures accurate reconstruction of voxel blocks and textures from the encoded data.
• Seamless mesh representation: The reconstructed mesh representation of the object allows for smooth visualization and manipulation of three-dimensional objects.

Benefits

The use of this technology offers several benefits, including:

• Reduced storage requirements: By encoding voxel blocks and textures, the system reduces the amount of storage space required for three-dimensional representations.
• Efficient transmission: The encoded data can be transmitted more efficiently, enabling faster transfer of three-dimensional models over networks.
• Improved visualization: The reconstructed mesh representation provides a visually appealing and accurate representation of the object, enhancing the user experience.

Original Abstract Submitted

Systems and methods are directed to encoding and/or decoding of the textures/geometry of a three-dimensional volumetric representation. An encoding computing system can obtain voxel blocks from a three-dimensional volumetric representation of an object. The encoding computing system can encode voxel blocks with a machine-learned voxel encoding model to obtain encoded voxel blocks. The encoding computing system can decode the encoded voxel blocks with a machine-learned voxel decoding model to obtain reconstructed voxel blocks. The encoding computing system can generate a reconstructed mesh representation of the object based at least in part on the one or more reconstructed voxel blocks. The encoding computing system can encode textures associated with the voxel blocks according to an encoding scheme and based at least in part on the reconstructed mesh representation of the object to obtain encoded textures.