Neural Scene Designer: Self-Styled Semantic Image Manipulation
Published in IEEE Transactions on Image Processing, Vol. 34, pp. 6577–6588, 2025 · CCF-A, 2025
NSD uses intra-image style consistency as a self-supervised signal and a dual parallel cross-attention design for disentangled text vs. style manipulation in diffusion models, with PSRL for self-style representation learning.
Recommended citation: Jianman Lin, Tianshui Chen, Chunmei Qing, Zhijing Yang, Shuangping Huang, Yuheng Ren, Liang Lin. Neural Scene Designer: Self-Styled Semantic Image Manipulation. IEEE TIP, 34, 6577–6588, 2025.
Download Paper