VR Art Creation Based on Image Processing Technology in New Media Environment
Description
This study aims to develop a VR artistic creation system with immersive and intelligent interaction, thus contributing to reconstructing and optimizing three-dimensional (3D) artistic images in VR scenes. In the context of new media, this study discusses how to use image processing technology to solve the limitations of traditional graphic art creation in realizing users' deep immersion and interactive experience given the revolutionary demand brought by VR technology for artistic creation. Through a comprehensive analysis of the existing VR environment, a VR artistic image creation system is meticulously constructed, with its focal point being the artistic images within the VR realm and drawing data from the WikiArt dataset. Firstly, a novel 3D artistic image reconstruction method is introduced, rooted in weakly supervised learning. Leveraging the power of two-dimensional (2D) self-attention fusion (SAF) effectively harnesses 2D artistic images to enhance the process of 3D artistic image reconstruction, yielding significant optimizations in the learning outcomes of 3D models. Secondly, VR artistic images undergo refinement and enhancement through image processing technologies, encompassing image rendering, lighting adjustments, and shadow processing. These refinements are instituted to augment artistic images' realism and visual impact within VR scenes. Finally, the performance of the proposed VR artistic image creation system is rigorously assessed within the VR environment. The outcomes of this evaluation indicate that integrating 2D artistic images with 2D SAF successfully facilitates the generation of 3D artistic images within VR scenes. Moreover, in comparison to existing algorithms, the system presented here exhibits noteworthy improvements in accuracy (approximately 15%), completeness (around 7%), and overall performance (about 10%). Additionally, the reconstruction time for 3D artistic images is notably reduced by approximately 15 seconds. Therefore, this study demonstrates that the proposed VR artistic image creation system can efficiently and accurately solve the 3D reconstruction problem of 2D artistic images. It has broad application prospects and social values for the intelligent development of VR artistic creation in the new media environment.