Multi-style transfer and fusion of image’s regions based on attention mechanism and instance segmentation
作者：
Highlights：
• Propose a simplified multi-style transfer methods based on attention mechanism.
• Can make the stylized artworks effectively preserve the original salience semantic and visual features of the content images.
• Can realize the regional stylization with natural transition among the regions of the stylized works.
• Define a effective metric for evaluating the performance of a stylization model in terms of semantic preservation.
• Can better emphasize the theme of original contents and enrich the visual effects with more diverse styles.
摘要
•Propose a simplified multi-style transfer methods based on attention mechanism.•Can make the stylized artworks effectively preserve the original salience semantic and visual features of the content images.•Can realize the regional stylization with natural transition among the regions of the stylized works.•Define a effective metric for evaluating the performance of a stylization model in terms of semantic preservation.•Can better emphasize the theme of original contents and enrich the visual effects with more diverse styles.
论文关键词：Multi-style transfer,Attention mechanism,Regional stylization,Instance segmentation,Poisson fusion
论文评审过程：Received 21 June 2021, Revised 26 August 2022, Accepted 23 September 2022, Available online 28 September 2022, Version of Record 17 October 2022.
论文官网地址：https://doi.org/10.1016/j.image.2022.116871
