Metaverse-oriented Multi-task Learning Algorithms in Animation Scene Design

Authors

  • Zhuoyi Song Jiamusi University, Jiamusi 154007, Heilongjiang, China
  • Xinyu Guo Jiamusi University, Jiamusi 154007, Heilongjiang, China
  • Yu Gu Jiamusi University, Jiamusi 154007, Heilongjiang, China

Abstract

In response to the problems of poor efficiency in animation scene design and the inadequate immersive experience of users , this study focuses on the metaverse and takes the animation Spirited Away as the object, applying multi-task learning algorithms to examine animation scene design. Firstly, a multi-task learning framework based on generative adversarial network (GAN) is constructed. The deep convolution generative adversarial networks method (DCGAN) is used as the basic architecture of the generator, and the convolutional neural network (CNN) is used as the basic architecture of the discriminator to model and generate the environment of the animation scene. Then, physically-based rendering (PBR) technology is combined to simulate the propagation and reflection of light in the scene for lighting modeling. Percentage closer filtering (PCF) technology is used to process soft shadows so as to enhance the realism of the animation scenes. The Monte Carlo path tracing algorithm is used to comprehensively calculate the global lighting effects of various areas in the scene. Finally, the Tone Mapping algorithm is used for color space conversion and mapping processing, and efficient rendering and realistic effects are demonstrated through deferred rendering and dynamic lighting processing techniques, resulting in the final dynamic scene generation rendering. The experiment is divided into two tasks: dynamic environment modeling and light and shadow effect design and optimization. MT-DCGAN (multi-task deep convolutional generative adversarial network) achieves a design efficiency of 95.73%, with an improvement of 13.49% compared to that of DCGAN, and a user’s attention concentration of 93.36%. The combination of multi-task learning algorithms in animation scene design greatly improves design efficiency and user experience, and enhances user satisfaction with animation scenes, thereby contributing an important innovative approach to the animation industry.

Keywords: Metaverse direction; animation scene design; multi-task learning algorithm; Spirited Away; immersive experience

Cite As

Z. Song, X. Guo, Y. Gu, "Metaverse-oriented Multi-task Learning Algorithms in Animation Scene Design", Engineering Intelligent Systems, vol. 34 no. 5, pp. 627-639, 2026

Downloads

Published

2026-09-01

Issue

Vol. 34 No. 5 (2026): International Journal of Engineering Intelligent Systems

Section

Article

License

The submission of a paper implies that, if accepted for publication, it will not be published elsewhere in the same form, in any language, without the prior consent of the publisher. Before publication, authors are requested to assign copyright to CRL Publishing Ltd. This allows CRL to sanction photocopying, and to authorize the reprinting of issues or volumes according to demand. Authors' traditional rights will not be jeopardized by assigning Copyright in this way, as they retain the right to reuse the material following publication, and to veto third-party publication.