arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2601.11522cs.CV
收藏

UniX: 统一自回归与扩散以实现胸部X光理解与生成

UniX: Unifying Autoregression and Diffusion for Chest X-Ray Understanding and Generation

Wuhan University(武汉大学);Huazhong University of Science and Technology(华中科技大学);Nanyang Technological University(南洋理工大学)

Ruiheng Zhang, Jingfeng Yao, Huangxuan Zhao, Hao Yan, Xiao He, Lei Chen, Zhou Wei, Yong Luo, Zengmao Wang, Lefei Zhang, Dacheng Tao, Bo Du

AI总结:

UniX通过统一自回归与扩散模型,实现了胸部X光的高效理解和生成,显著提升了性能并减少了参数使用。

摘要:

尽管近年来取得了进展,但医学基础模型仍然难以统一视觉理解和生成,因为这些任务本质上存在冲突的目标:语义抽象与像素级重建。现有方法通常基于参数共享的自回归架构,常常导致在一项或两项任务上的性能妥协。为了解决这一问题,我们提出了UniX,一种新一代统一的医学基础模型,用于胸部X光的理解和生成。UniX将两项任务解耦为一个自回归分支用于理解,一个扩散分支用于高质量生成。关键的是,引入了跨模态自注意力机制,以动态引导生成过程,利用理解特征。结合严谨的数据清洗流程和多阶段训练策略,这种架构使任务之间能够实现协同合作,同时利用扩散模型的优势实现更优的生成效果。在两个具有代表性的基准测试中,UniX在理解性能(Micro-F1)上实现了46.1%的提升,在生成质量(FD-RadDino)上获得了24.2%的提升,仅使用LLM-CXR参数的四分之一。通过在性能上与专门任务模型相当,我们的工作建立了一种可扩展的协同医学图像理解和生成范式。代码和模型可在https://github.com/ZrH42/UniX上获取。

英文摘要:

Despite recent progress, medical foundation models still struggle to unify visual understanding and generation, as these tasks have inherently conflicting goals: semantic abstraction versus pixel-level reconstruction. Existing approaches, typically based on parameter-shared autoregressive architectures, frequently lead to compromised performance in one or both tasks. To address this, we present UniX, a next-generation unified medical foundation model for chest X-ray understanding and generation. UniX decouples the two tasks into an autoregressive branch for understanding and a diffusion branch for high-fidelity generation. Crucially, a cross-modal self-attention mechanism is introduced to dynamically guide the generation process with understanding features. Coupled with a rigorous data cleaning pipeline and a multi-stage training strategy, this architecture enables synergistic collaboration between tasks while leveraging the strengths of diffusion models for superior generation. On two representative benchmarks, UniX achieves a 46.1% improvement in understanding performance (Micro-F1) and a 24.2% gain in generation quality (FD-RadDino), using only a quarter of the parameters of LLM-CXR. By achieving performance on par with task-specific models, our work establishes a scalable paradigm for synergistic medical image understanding and generation. Codes and models are available at https://github.com/ZrH42/UniX.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑