4 papers · 1 filter
CanvasAgent: Enabling Complex Image Creation and Editing via Visual Tool Orchestration
Hairui Zhu, Yiying Yang, Tengjin Weng +5
Complex image creation and editing often require more than a single generation or editing model. A user request may involve synthesizing images, localizing objects, segmenting regi…
OddGridBench: Exposing the Lack of Fine-Grained Visual Discrepancy Sensitivity in Multimodal Large Language Models
Tengjin Weng, Wenhao Jiang, Jingyi Wang +3
Multimodal large language models (MLLMs) have achieved remarkable performance across a wide range of vision language tasks. However, their ability in low-level visual perception, p…
UniViTAR: Unified Vision Transformer with Native Resolution
Limeng Qiao, Yiyang Gan, Bairui Wang +4
Conventional Vision Transformer simplifies visual modeling by standardizing input resolutions, often disregarding the variability of natural visual data and compromising spatial-co…
TIE: Revolutionizing Text-based Image Editing for Complex-Prompt Following and High-Fidelity Editing
Xinyu Zhang, Mengxue Kang, Fei Wei +3
As the field of image generation rapidly advances, traditional diffusion models and those integrated with multimodal large language models (LLMs) still encounter limitations in int…