Research on the Implementation of Chinese Painting Artistic Concepts in Computer-Assisted Film and Television Animation Creation

View at publisher Summarise with VIRAT
Authors: Hua Zhu, Chi Zhang
Cited by: 0
References: 5

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

To address the difficulty of existing style transfer algorithms in accurately transferring the artistic conception of Chinese ink paintings, a multi-feature fusion-based Chinese ink painting artistic conception transfer fusion algorithm is proposed. This algorithm constructs a three-step framework: feature layer extraction - dynamic weight fusion - physical simulation enhancement. It improves the VGG19 network to extract brush and ink features, combines it with a spatial attention model to capture composition and white space features, extracts animation dynamic features using optical flow and contour constraints, dynamically assigns feature weights based on information evaluation, and introduces a physical model of ink diffusion to optimize the generation effect. The experiments are conducted using CycleGAN and StyleGAN2 as benchmarks, and are validated on a dataset consisting of 400 Chinese paintings for training and 2800 animation frames. Results show that the proposed algorithm achieved an average subjective evaluation score of 8.52, representing improvements of 40.8% and 28.1% over CycleGAN (6.05) and StyleGAN2 (6.65), respectively. Objectively, the algorithm achieved average PSNR of 32.5dB, SSIM of 0.89, and FSIM of 0.92, respectively, representing improvements of 9.4%-17.9% over the baseline algorithms. The algorithm converged after 12,000 iterations, 25% faster than CycleGAN, effectively achieving the precise coupling of the artistic conception of Chinese ink painting with film and television animation.

Fields of Study & Subjects

Generative Adversarial Networks and Image Synthesis Digital Media and Visual Art Computer Graphics and Visualization Techniques

Publication details

Journal / source
IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
Pages
535–540
Publication date
28 January 2026
Language
EN