用金字塔位置编码和强化学习，实现快速高质量图像风格迁移。
PyramidStyler: Transformer-Based Neural Style Transfer with Pyramidal Positional Encoding and Reinforcement Learning
- 引入金字塔位置编码，高效捕捉多尺度细节与全局结构。
- 训练4000轮后内容损失降为2.07，风格损失降至0.86，推理仅需1.39秒。
- 结合强化学习可进一步提升效果，适合媒体设计等实时应用。
神经风格迁移（NST）已从Gatys等人（2015）基于CNN的算法演进为支持AI生成艺术图像的技术。然而，现有基于CNN或Transformer的模型在处理复杂风格和高分辨率输入时难以高效扩展。我们提出PyramidStyler，一种采用金字塔位置编码（PPE）的Transformer框架：该编码方式分层、多尺度地捕捉局部细节与全局上下文，同时降低计算开销。我们进一步引入强化学习动态优化风格化过程，加速收敛。在Microsoft COCO和WikiArt数据集上训练后，PyramidStyler在4000轮迭代中将内容损失降低62.6%（至2.07），风格损失降低57.4%（至0.86），推理时间仅为1.39秒；使用强化学习后，内容损失进一步降至2.03，风格损失降至0.75，仅带来轻微速度代价（1.40秒）。结果表明，该方法实现了实时且高质量的艺术渲染，适用于媒体与设计等领域。
原文摘要 · Abstract (English)
Neural Style Transfer (NST) has evolved from Gatys et al.'s (2015) CNN-based algorithm, enabling AI-driven artistic image synthesis. However, existing CNN and transformer-based models struggle to scale efficiently to complex styles and high-resolution inputs. We introduce PyramidStyler, a transformer framework with Pyramidal Positional Encoding (PPE): a hierarchical, multi-scale encoding that captures both local details and global context while reducing computational load. We further incorporate reinforcement learning to dynamically optimize stylization, accelerating convergence. Trained on Microsoft COCO and WikiArt, PyramidStyler reduces content loss by 62.6% (to 2.07) and style loss by 57.4% (to 0.86) after 4000 epochs--achieving 1.39 s inference--and yields further improvements (content 2.03; style 0.75) with minimal speed penalty (1.40 s) when using RL. These results demonstrate real-time, high-quality artistic rendering, with broad applications in media and design.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。