arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2403.07214cs.CV
收藏

文本到图像扩散模型是出色的草图-照片匹配器

Text-to-Image Diffusion Models are Great Sketch-Photo Matchmakers

University of Surrey(萨里大学);iFlyTek-Surrey Joint Research Centre on Artificial Intelligence(科大讯飞-萨里人工智能联合研究中心)

Subhadeep Koley, Ayan Kumar Bhunia, Aneeshan Sain, Pinaki Nath Chowdhury, Tao Xiang, Yi-Zhe Song

AI总结:

本文首次探索文本到图像扩散模型在零样本基于草图的图像检索(ZS-SBIR)中的应用,提出通过选择最优特征层并结合视觉与文本提示的策略,显著提升了跨模态检索性能。

摘要:

本文首次探索了文本到图像扩散模型在零样本基于草图的图像检索(ZS-SBIR)中的应用。我们强调了一项关键发现:文本到图像扩散模型能够无缝弥合草图与照片之间的差距。这种能力由其强大的跨模态能力和形状偏差所支撑,这一发现通过我们的初步研究得到了证实。为了有效利用预训练的扩散模型,我们引入了一种简单而强大的策略,重点关注两个关键方面:选择最优特征层和利用视觉与文本提示。对于前者,我们识别出哪些层包含最丰富的信息,并且最适合特定的检索需求(类别级或细粒级)。然后,我们采用视觉和文本提示来引导模型的特征提取过程,使其能够生成更具判别性和上下文相关的跨模态表示。在多个基准数据集上的大量实验验证了显著的性能提升。

英文摘要:

This paper, for the first time, explores text-to-image diffusion models for Zero-Shot Sketch-based Image Retrieval (ZS-SBIR). We highlight a pivotal discovery: the capacity of text-to-image diffusion models to seamlessly bridge the gap between sketches and photos. This proficiency is underpinned by their robust cross-modal capabilities and shape bias, findings that are substantiated through our pilot studies. In order to harness pre-trained diffusion models effectively, we introduce a straightforward yet powerful strategy focused on two key aspects: selecting optimal feature layers and utilising visual and textual prompts. For the former, we identify which layers are most enriched with information and are best suited for the specific retrieval requirements (category-level or fine-grained). Then we employ visual and textual prompts to guide the model's feature extraction process, enabling it to generate more discriminative and contextually relevant cross-modal representations. Extensive experiments on several benchmark datasets validate significant performance improvements.

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。

↑