Apparatus, system and method for data generation

A method of training a system to learn a function for adapting data of an input polygonal mesh to generate data of an output polygonal mesh, comprising:

• accessing a training data set comprising, for a reference polygonal mesh and a second polygonal mesh each comprising a plurality of vertices for representing a same surface, data indicative of a vertex count of each of the reference and the second polygonal meshes and a value of a quality parameter for each of the reference and the second polygonal meshes, the quality parameter for a given polygonal mesh indicative of a degree of difference between a representation of the given polygonal mesh and a representation of the reference polygonal mesh, the vertex count of the second polygonal mesh being greater than the vertex count of the reference polygonal mesh, wherein
• the training data set comprises data of the reference and second polygonal meshes for a plurality of different surfaces;
• analysing the data for the second polygonal meshes and the reference polygonal meshes of the training data set to determine a candidate function for removing a plurality of vertices from the second polygonal meshes to generate the reference polygonal meshes;
• obtaining data of an input polygonal mesh comprising a first number of polygon surfaces, each polygon surface defined by a plurality of vertices;
• adapting the data of the input polygonal mesh by removing a plurality of vertices from the input polygonal mesh according to the candidate function to generate data of an output polygonal mesh;
• comparing a representation of the output polygonal mesh with a representation of one or more of the reference polygonal meshes of the training data set;
• assigning a value of a quality parameter for the output polygonal mesh based on the comparison of the representations;
• and
• adapting the candidate function in dependence upon the value of the quality parameter for the output polygonal mesh.

A method of training a system to learn a function for adapting data of an input polygonal mesh to generate data of an output polygonal mesh, comprising:

• accessing a training data set comprising data, for a reference, second and third polygonal mesh each comprising a plurality of vertices for representing a same surface in a content, data indicative of a vertex count for each of the reference, second and third polygonal meshes and a value of a quality parameter for each of the reference, second and third polygonal meshes, the quality parameter for a given polygonal mesh indicative of a degree of difference between a representation of the given polygonal mesh and a representation of the reference polygonal mesh, the vertex count of the second polygonal mesh being greater than the vertex count of the reference polygonal mesh and the vertex count of the reference polygonal mesh being greater than the vertex count of the third polygonal mesh, wherein
• the training data set comprises data of the reference, second and third polygonal meshes for a plurality of different surfaces;
• obtaining data of an input polygonal mesh comprising a first number of polygon surfaces, each polygon surface defined by a plurality of vertices;
• assigning a value of the quality parameter for the input polygonal mesh based on a comparison of a representation of the input polygonal mesh with a representation of one or more of the reference polygonal meshes of the training data set;
• removing a plurality of vertices from the input polygonal mesh according to a candidate function to generate data of an output polygonal mesh;
• comparing a representation of the output polygonal mesh with a representation of the input polygonal mesh and determining a difference between a value of the quality parameter for the output polygonal mesh and the value of the quality parameter for the input polygonal mesh based on the comparison of the representations;
• and
• adapting the candidate function for removing the plurality of vertices from the input polygonal mesh in dependence upon the difference in the value of the quality parameter for the output polygonal mesh and the input polygonal mesh.

A non-transitory, computer-readable storage medium containing computer software which, when executed by a computer, causes the computer to tain a system to learn a function for adapting data of an input polygonal mesh to generate data of an output polygonal mesh by carrying out actions, comprising:

• accessing a training data set comprising data, for a reference, second and third polygonal mesh each comprising a plurality of vertices for representing a same surface in a content, data indicative of a vertex count for each of the reference, second and third polygonal meshes and a value of a quality parameter for each of the reference, second and third polygonal meshes, the quality parameter for a given polygonal mesh indicative of a degree of difference between a representation of the given polygonal mesh and a representation of the reference polygonal mesh, the vertex count of the second polygonal mesh being greater than the vertex count of the reference polygonal mesh and the vertex count of the reference polygonal mesh being greater than the vertex count of the third polygonal mesh, wherein
• training data set comprises data of the reference, second and third polygonal meshes for a plurality of different surfaces;
• obtaining data of an input polygonal mesh comprising a first number of polygon surfaces, each polygon surface defined by a plurality of vertices;
• assigning a value of the quality parameter for the input polygonal mesh based on a comparison of a representation of the input polygonal mesh with a representation of one or more of the reference polygonal meshes of the training data set
• removing a plurality of vertices from the input polygonal mesh according to a candidate function to generate data of an output polygonal mesh;
• comparing a representation of the output polygonal mesh with a representation of the input polygonal mesh and determining a difference between a value of the quality parameter for the output polygonal mesh and the value of the quality parameter for the input polygonal mesh based on the comparison of the representations;
• and
• adapting the candidate function for removing the plurality of vertices from the input polygonal mesh in dependence upon the difference in the value of the quality parameter for the output polygonal mesh and the input polygonal mesh.

A system according to claim 17, further comprising:

• apparatus to generate a data set for training a processor to generate data of a polygonal mesh by adapting data of an input polygonal mesh, the apparatus comprising circuitry to:
• obtain data of a plurality of reference polygonal meshes for use in rendering a content, each reference polygonal mesh comprising a plurality of vertices for representing a surface in the content;
• determine a vertex count of each of the reference polygonal meshes;
• generate a second polygonal mesh by adding a predetermined number of vertices to the reference polygonal mesh;
• compare a representation of the second polygonal mesh with a representation of the reference polygonal mesh;
• assign a first value a quality parameter to the reference polygonal mesh and to assign a second value of the quality parameter to the second polygonal mesh based on the comparison of the representations, the value of the quality parameter for a given polygonal mesh indicative of a degree of difference between a representation of the given polygonal mesh and a representation of the reference polygonal mesh;
• and
• generate a data set comprising data indicative of the vertex count for the reference and second polygonal meshes and the value of the quality parameters for the reference and second polygonal meshes.

A system configured to learn a function for adapting data of an input polygonal mesh to generate data of an output polygonal mesh, comprising circuitry configured to:

• access a training data set comprising, for a reference polygonal mesh and a second polygonal mesh each comprising a plurality of vertices for representing a same surface, data indicative of a vertex count of each of the reference and the second polygonal meshes and a value of a quality parameter for each of the reference and the second polygonal meshes, the quality parameter for a given polygonal mesh indicative of a degree of difference between a representation of the given polygonal mesh and a representation of the reference polygonal mesh, the vertex count of the second polygonal mesh being greater than the vertex count of the reference polygonal mesh, wherein
• the training data set comprises data of the reference and second polygonal meshes for a plurality of different surfaces;
• analyse the data for the second polygonal meshes and the reference polygonal meshes of the training data set to determine a candidate function for removing a plurality of vertices from the second polygonal meshes to generate the reference polygonal meshes;
• obtain data of an input polygonal mesh comprising a first number of polygon surfaces, each polygon surface defined by a plurality of vertices;
• adapt the data of the input polygonal mesh by removing a plurality of vertices from the input polygonal mesh according to the candidate function to generate data of an output polygonal mesh;
• compare a representation of the output polygonal mesh with a representation of one or more of the reference polygonal meshes of the training data set;
• assign a value of a quality parameter for the output polygonal mesh based on the comparison of the representations;
• and
• adapt the candidate function in dependence upon the value of the quality parameter for the output polygonal mesh.

A system configured to learn a function for adapting data of an input polygonal mesh to generate data of an output polygonal mesh, comprising circuity configured to:

• access a training data set comprising data, for a reference, second and third polygonal mesh each comprising a plurality of vertices for representing a same surface in a content, data indicative of a vertex count for each of the reference, second and third polygonal meshes and a value of a quality parameter for each of the reference, second and third polygonal meshes, the quality parameter for a given polygonal mesh indicative of a degree of difference between a representation of the given polygonal mesh and a representation of the reference polygonal mesh, the vertex count of the second polygonal mesh being greater than the vertex count of the reference polygonal mesh and the vertex count of the reference polygonal mesh being greater than the vertex count of the third polygonal mesh, wherein
• the training data set comprises data of the reference, second and third polygonal meshes for a plurality of different surfaces;
• obtain data of an input polygonal mesh comprising a first number of polygon surfaces, each polygon surface defined by a plurality of vertices;
• assign a value of the quality parameter for the input polygonal mesh based on a comparison of a representation of the input polygonal mesh with a representation of one or more of the reference polygonal meshes of the training data set;
• remove a plurality of vertices from the input polygonal mesh according to a candidate function to generate data of an output polygonal mesh;
• compare a representation of the output polygonal mesh with a representation of the input polygonal mesh and determining a difference between a value of the quality parameter for the output polygonal mesh and the value of the quality parameter for the input polygonal mesh based on the comparison of the representations;
• and
• adapt the candidate function for removing the plurality of vertices from the input polygonal mesh in dependence upon the difference in the value of the quality parameter for the output polygonal mesh and the input polygonal mesh.