2 papers
cs.CV2026
DICT: Data Injection and Contrastive Trajectory Refinement for Conditional Image Generation with Diffusion Models
Chunnan Shang, Xin Zhang, Zhizhong Wang +1
Diffusion models have become a dominant paradigm for conditional image generation, yet existing approaches generally follow two directions: task-specific designs that can improve p…
cs.CV2025
SCSA: A Plug-and-Play Semantic Continuous-Sparse Attention for Arbitrary Semantic Style Transfer
Chunnan Shang, Zhizhong Wang, Hongwei Wang +1
Attention-based arbitrary style transfer methods, including CNN-based, Transformer-based, and Diffusion-based, have flourished and produced high-quality stylized images. However, t…