Research on Denoising Technology for Yimeng Red Films Based on Swin Transformer

This research introduces a novel method for denoising Yimeng red films by integrating Swin Transformer and Generative Adversarial Network (GAN). The Swin Transformer effectively extracts both local and global features of video frames through its hierarchical structure and window attention mechanism. Meanwhile, the GAN enhances the realism and detail preservation of the denoised videos through adversarial training. Experimental results show that the proposed method achieves a PSNR value of 36.47 and an SSIM value of 0.940 in the YOUKUVESR dataset, outperforming other advanced algorithms. This provides a new solution for old film restoration and contributes to the preservation of cultural heritage.