2026· European Conference on Computer Vision· pp. 518-536· 0 citations· 64 references
Computer Science
TL;DR
A novel self-supervised framework that enables granular control over image generation through a visual abstraction set that provides a richer, more flexible paradigm for creative design compared to state-of-the-art baselines across diverse styles and compositions is introduced.
Editable Visual Design is proposed, a new paradigm driven by a Coding Agent that achieves both refined aesthetics and production-grade editability and faithfully reproduces the creative and reasoning trajectory akin to that of professional human designers.
Jun-Yan Ye, Wei Liu, Dongzhi Jiang et al.· 0 citations
Pose-Star++ is proposed, a training-free, plug-and-play framework that introduces two core innovations: an LVLM-based Understanding Module that shifts from word- to sentence-level semantic-visual comprehension, eliminating cumbersome instruction pre-parsing and enabling robust understanding of complex natural language.
Yuran Dong, Bo Du, Mang Ye· IEEE Transactions on Pattern...· 0 citations
An automated pipeline leveraging Multimodal Large Language Models (MLLMs) is developed to synthesize comprehensive quadruplets comprising original images, local geometric sketches, semantic instructions, and corresponding edited images, which uniquely enables collaborative spatial-semantic learning.
Weixin Ye, Wei Wang, Hong-Guang Zhu et al.· 0 citations
This work analyzes the semantic structure of abstract art through large-scale embedding visualization, uncovering how perceptual relationships organize artistic meaning, and establishes benchmark tasks for classification, cross-modal retrieval, and text-to-image generation to evaluate how AI models perceive and reprodu...
Hao-Wei Zhang, Yuanpei Zhao, Ji-Zhe Zhou et al.· 0 citations
Paint-Anything is presented, which learns a shared hex-prompt interface for generation and editing through object-level color supervision, and introduces Any Color Benchmark (ACBench), comprising ACBench-T2I and ACBench-Edit, to measure object-level hex color fidelity across both tasks.
Ji Xie, Dewei Zhou, Xin-Yu Huang et al.· 0 citations
StyleSmith is proposed, a diffusion framework for high-fidelity signature style transfer that supports local style transfer, texture transfer, and style-guided text-to-image generation, demonstrating strong academic innovation and industrial value.
