Search

CN-121982272-A - Object processing method, data processing method and object processing model training method

CN121982272ACN 121982272 ACN121982272 ACN 121982272ACN-121982272-A

Abstract

The specification provides an object processing method, a data processing method and an object processing model training method, wherein the object processing method comprises the steps of determining target bone data corresponding to a plurality of object parts of a target object, inputting the target bone data into an object processing model to obtain associated bone data corresponding to an associated object of the target object, wherein the associated object is an object worn by the target object, the associated bone data is determined according to sub-associated bone data corresponding to at least one sub-associated object, the sub-associated bone data is determined according to fusion bone characteristics corresponding to an object part set, the fusion bone characteristics are determined according to the target bone data corresponding to the object parts contained in the object part set, and the associated object is generated according to the associated bone data. The object generating method can be widely applied to the fields of digital cultural product making software, digital cultural creative design and the like in the digital creative technology.

Inventors

  • LI YANG

Assignees

  • 珠海金山数字网络科技有限公司

Dates

Publication Date
20260505
Application Date
20260129

Claims (14)

  1. 1. An object processing method, comprising: determining target bone data corresponding to a plurality of object parts of a target object; Inputting the target bone data into an object processing model to obtain associated bone data corresponding to an associated object of the target object, wherein the associated object is an object worn by the target object, the associated bone data is determined according to sub-associated bone data corresponding to at least one sub-associated object, the sub-associated bone data is determined according to fusion bone characteristics corresponding to an object part set, and the fusion bone characteristics are determined according to target bone data corresponding to an object part contained in the object part set; and generating the association object according to the association bone data.
  2. 2. The method according to claim 1, wherein inputting the target bone data into an object processing model to obtain associated bone data corresponding to an associated object of the target object comprises: inputting the target bone data into an object processing model, and performing feature coding on the target bone data in the object processing model to obtain target bone features corresponding to the plurality of object parts; grouping the plurality of object parts to obtain at least one object part set; Determining sub-association bone data corresponding to at least one sub-association object according to target bone characteristics corresponding to the object parts contained in each object part set, wherein the sub-association objects correspond to the object part sets and have a position association relationship; and determining associated bone data corresponding to the associated object according to the at least one sub-associated bone data.
  3. 3. The method according to claim 2, wherein determining sub-associated bone data corresponding to at least one sub-associated object according to the target bone features corresponding to the object parts included in the respective object part sets comprises: Fusing target bone characteristics corresponding to the object parts contained in each object part set to obtain fused bone characteristics corresponding to each object part set; And determining sub-associated bone data corresponding to at least one sub-associated object based on the fused bone features and the target bone features corresponding to the at least one object part set.
  4. 4. The method of claim 1, wherein determining target bone data corresponding to a plurality of object locations of a target object comprises: determining target action data corresponding to a target object; And extracting skeleton rotation data corresponding to a plurality of object parts of the target object according to the target action data to serve as the target skeleton data.
  5. 5. The method according to claim 4, wherein extracting bone rotation data corresponding to a plurality of object parts of the target object as the target bone data according to the target motion data, further comprises: Extracting skeleton rotation data corresponding to a plurality of object parts of the target object according to the target action data; and based on the six-dimensional vector, carrying out rotary encoding on the bone rotation data corresponding to the plurality of object parts, and obtaining converted bone rotation data serving as the target bone data.
  6. 6. The method of any of claims 1-5, wherein generating the association object from the association bone data comprises: and performing skin binding on the associated bone data to generate the associated object.
  7. 7. The method according to any one of claims 1 to 5, wherein before inputting the target bone data into the object processing model to obtain associated bone data corresponding to an associated object of the target object, the method further comprises: Acquiring sample target bone data corresponding to a plurality of object parts of a sample target object and sample associated bone data corresponding to a sample associated object; And training the object processing model according to the sample target bone data and the sample associated bone data until reaching a training stopping condition, and obtaining the trained object processing model.
  8. 8. The method of claim 7, wherein the acquiring sample target bone data corresponding to a plurality of subject sites of a sample target subject comprises: Determining sample target action data corresponding to a sample target object; extracting sample skeleton rotation data corresponding to a plurality of object parts of the sample target object according to the sample target action data; And carrying out rotary encoding on the sample skeleton rotary data corresponding to the plurality of object parts based on the six-dimensional vector, and obtaining converted sample skeleton rotary data serving as the sample target skeleton data.
  9. 9. The method according to claim 8, wherein the rotating encoding the sample bone rotation data corresponding to the plurality of object sites based on the six-dimensional vector, before obtaining the converted sample bone rotation data as the sample target bone data, further comprises: Adding preset sample noise to the sample bone rotation data corresponding to the plurality of object parts to obtain noisy sample bone rotation data; The rotating encoding of the sample bone rotation data corresponding to the plurality of object parts based on the six-dimensional vector, obtaining converted sample bone rotation data as the sample target bone data, includes: and carrying out rotary coding on the noise-added sample skeleton rotation data based on the six-dimensional vector to obtain converted noise-added sample skeleton rotation data serving as the sample target skeleton data.
  10. 10. A method of data processing, comprising: determining target skeleton data corresponding to a plurality of character parts of a target virtual three-dimensional character; Inputting the target bone data into an object processing model to obtain associated bone data corresponding to a clothing object of the target virtual three-dimensional character, wherein the associated bone data is determined according to sub-associated bone data corresponding to at least one sub-clothing object, the sub-associated bone data is determined according to fusion bone characteristics corresponding to a character part set, and the fusion bone characteristics are determined according to the target bone data corresponding to the character part contained in the character part set; and generating the clothing object according to the associated bone data.
  11. 11. A method of training an object processing model, comprising: Acquiring sample target bone data corresponding to a plurality of object parts of a sample target object and sample associated bone data corresponding to a sample associated object, wherein the sample associated object is an object worn by the sample target object; Inputting the sample target bone data into an object processing model to obtain prediction associated bone data corresponding to the sample associated object, wherein the prediction associated bone data is determined according to sample sub-associated bone data corresponding to at least one sample sub-associated object, the sample sub-associated bone data is determined according to sample fusion bone characteristics corresponding to an object part set, and the sample fusion bone characteristics are determined according to sample target bone data corresponding to an object part contained in the object part set; And training the object processing model according to the sample associated bone data and the prediction associated bone data until a training stopping condition is reached, so as to obtain a trained object processing model.
  12. 12. A computing device, comprising: A memory and a processor; the memory is adapted to store a computer program/instruction, the processor being adapted to execute the computer program/instruction, which when executed by the processor, performs the steps of the method of any of claims 1 to 11.
  13. 13. A computer readable storage medium storing a computer program/instruction, which when executed by a processor performs the steps of the method of any one of claims 1 to 11.
  14. 14. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method of any one of claims 1 to 11.

Description

Object processing method, data processing method and object processing model training method Technical Field The present disclosure relates to the field of digital cultural creative activities and the field of artificial intelligence technologies, and in particular, to an object processing method, a data processing method, and an object processing model training method. Background With the development of game technology, in the production of 3D characters (such as game characters, movie and television animation characters and virtual persons), clothing is a key element for shaping the characters, and by wearing various clothing on the 3D characters, the reality and visual expressive force on the 3D characters are realized. In general, in order to avoid the wearing condition of the clothing, the skeleton of the clothing can be used as a motion control fulcrum to realize the cooperative linkage of the actions of the clothing and the character. However, when creating a clothing bone currently, it is generally necessary to manually set the position of the clothing bone, which consumes manpower and time, and relies heavily on manual experience, resulting in low efficiency. Therefore, an effective solution is needed to solve the above problems. Disclosure of Invention In view of this, the present embodiment provides an object processing method. The present specification also relates to an object processing apparatus, a data processing method, a data processing apparatus, an object processing model training method, an object processing model training apparatus, a computing device, a computer-readable storage medium, and a computer program product, for solving the above-described problems in the prior art. According to a first aspect of embodiments of the present specification, there is provided an object processing method, including: determining target bone data corresponding to a plurality of object parts of a target object; Inputting the target bone data into an object processing model to obtain associated bone data corresponding to an associated object of the target object, wherein the associated object is an object worn by the target object, the associated bone data is determined according to sub-associated bone data corresponding to at least one sub-associated object, the sub-associated bone data is determined according to fusion bone characteristics corresponding to an object part set, and the fusion bone characteristics are determined according to target bone data corresponding to an object part contained in the object part set; and generating the association object according to the association bone data. According to a second aspect of embodiments of the present specification, there is provided an object processing apparatus comprising: the determining module is configured to determine target bone data corresponding to a plurality of object parts of a target object; the input module is configured to input the target bone data into an object processing model to obtain associated bone data corresponding to an associated object of the target object, wherein the associated object is an object worn by the target object, the associated bone data is determined according to sub-associated bone data corresponding to at least one sub-associated object, the sub-associated bone data is determined according to fusion bone characteristics corresponding to an object position set, and the fusion bone characteristics are determined according to target bone data corresponding to an object position contained in the object position set; and the generation module is configured to generate the association object according to the association skeleton data. According to a third aspect of embodiments of the present specification, there is provided a data processing method, comprising: determining target skeleton data corresponding to a plurality of character parts of a target virtual three-dimensional character; Inputting the target bone data into an object processing model to obtain associated bone data corresponding to a clothing object of the target virtual three-dimensional character, wherein the associated bone data is determined according to sub-associated bone data corresponding to at least one sub-clothing object, the sub-associated bone data is determined according to fusion bone characteristics corresponding to a character part set, and the fusion bone characteristics are determined according to the target bone data corresponding to the character part contained in the character part set; and generating the clothing object according to the associated bone data. According to a fourth aspect of embodiments of the present specification, there is provided a data processing apparatus comprising: The determining module is configured to determine target skeleton data corresponding to a plurality of character parts of a target virtual three-dimensional character; The input module is configured to input the target bone data into a