With the development of the digital age, online exhibitions of artworks are becoming increasingly popular, but visual effects problems frequently occur. About 70% of artworks have room for improvement in visual presentation, and the feedback rate of online exhibition audiences being dissatisfied with the visual effects is as high as 45%. To solve this problem, this paper proposes a Multimodal Learning Framework for Visual Enhancement of Digital Artworks (ArtFusionNet) for enhancing the visual effects of artworks based on multimodal learning. The algorithm integrates multimodal information such as color, line, light, and shadow through feature extraction, attention mechanism, multimodal fusion, and visual effect optimization modules. A comparison was made between deep learning-based picture super-resolution techniques and more conventional methods of color correction. The studies were carried out on several datasets, including WikiArt and ArtBench-10, and the results were quite interesting. The results show that the technique proposed in this research improves the color restoration of artworks by 40% and the richness of detail by 35%. Regarding color restoration of oil paintings, the average color difference is only 2.14, much lower than the comparison algorithm. This study enriches the theory of multimodal learning and visual effect optimization of artworks in computer vision, providing strong technical support for digital dissemination.
Ge et al. (Sat,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: