arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2506.01331cs.CV
收藏

超高分辨率图像合成:数据、方法与评估

Ultra-High-Resolution Image Synthesis: Data, Method and Evaluation

State Key Laboratory of Complex and Critical Software Environment(复杂与关键软件环境国家重点实验室);School of Computer Science and Engineering(计算机科学与工程学院);Beihang University(北京航空航天大学)

Jinjin Zhang, Qiuyu Huang, Junjie Liu, Xiefan Guo, Di Huang

AI总结:

本文构建 Aesthetic-4K 数据集并提出 Diffusion-4K 框架,通过 SC-VAE 与 WLF 技术直接生成超高分辨率 4K 图像,并引入 GLCM Score 等新指标实现全面评估。

摘要:

超高分辨率图像合成具有巨大潜力,但由于缺乏标准化基准和计算限制,仍然是一个未被充分探索的挑战。在本文中,我们构建了 Aesthetic-4K,这是一个精心整理的数据集,包含专门设计的训练和评估子集,旨在支持超高分辨率图像合成的全面研究。该数据集由高质量的 4K 图像组成,并配有由 GPT-4o 生成的描述性文本说明。此外,我们提出了 Diffusion-4K,这是一个用于直接生成超高分辨率图像的创新框架。我们的方法融合了尺度一致性变分自编码器(SC-VAE)和基于小波的潜在微调(WLF),旨在实现高效的视觉 token 压缩并捕获超高分辨率图像中的复杂细节,从而促进使用逼真 4K 数据进行直接训练。该方法适用于各种潜在扩散模型,并证明了其在合成高度精细的 4K 图像方面的有效性。此外,我们提出了新的评估指标,即 GLCM Score 和 Compression Ratio,用于评估局部图像块中的纹理丰富度和细节精细度,并结合 FID、Aesthetics 和 CLIPScore 等整体度量,从而实现对超高分辨率图像合成的全面、多维度评估。结果表明,Diffusion-4K 在超高分辨率图像合成中取得了令人瞩目的性能,尤其是在由最先进的大规模扩散模型(例如 Flux-12B)驱动时表现尤为突出。源代码已在 https://github.com/zhang0jhon/diffusion-4k 公开提供。

英文摘要:

Ultra-high-resolution image synthesis holds significant potential, yet remains an underexplored challenge due to the absence of standardized benchmarks and computational constraints. In this paper, we establish Aesthetic-4K, a meticulously curated dataset containing dedicated training and evaluation subsets specifically designed for comprehensive research on ultra-high-resolution image synthesis. This dataset consists of high-quality 4K images accompanied by descriptive captions generated by GPT-4o. Furthermore, we propose Diffusion-4K, an innovative framework for the direct generation of ultra-high-resolution images. Our approach incorporates the Scale Consistent Variational Auto-Encoder (SC-VAE) and Wavelet-based Latent Fine-tuning (WLF), which are designed for efficient visual token compression and the capture of intricate details in ultra-high-resolution images, thereby facilitating direct training with photorealistic 4K data. This method is applicable to various latent diffusion models and demonstrates its efficacy in synthesizing highly detailed 4K images. Additionally, we propose novel metrics, namely the GLCM Score and Compression Ratio, to assess the texture richness and fine details in local patches, in conjunction with holistic measures such as FID, Aesthetics, and CLIPScore, enabling a thorough and multifaceted evaluation of ultra-high-resolution image synthesis. Consequently, Diffusion-4K achieves impressive performance in ultra-high-resolution image synthesis, particularly when powered by state-of-the-art large-scale diffusion models (eg, Flux-12B). The source code is publicly available at https://github.com/zhang0jhon/diffusion-4k.

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。

↑