Image restoration model of wood-structured raw lacquer cultural relics based on diffusion generation network

Wooden cultural relics are highly prone to damage, such as peeling of the surface paint layer and fading of color due to environmental factors and improper preservation. Existing image restoration techniques have difficulty maintaining the authenticity of details and coherence of artistic styles. This study proposes an image restoration model for wooden structure lacquer cultural relics based on a diffusion generation network. The diffusion model in the previous step restores the basic structure and color distribution of the lacquer cultural relics images, and then the generation network in the subsequent step finely restores the details, textures, colors, etc. of the initially restored images, improving the visual clarity and fineness of the restored images. An image dataset of wooden lacquer cultural relics was constructed to support the training and verification of the model. The experimental results show that the model in this study is superior to existing algorithms for cultural relic image restoration in both subjective vision and objective indicators, providing new ideas and technical paths for the restoration of wooden lacquer cultural relics and promoting technological development in the field of ancient wooden building restoration.