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 2010.09641cs.MMcs.AI
收藏

DIME:一个跨模态检索模型可视化比较的在线工具

DIME: An Online Tool for the Visual Comparison of Cross-Modal Retrieval Models

University of California at Berkeley(加州大学伯克利分校);International Computer Science Institute(国际计算机科学研究所)

Tony Zhao, Jaeyoung Choi, Gerald Friedland

摘要:

跨模态检索依赖于准确的模型来检索跨模态(诸如图像、文本和视频)的查询的相关结果。在本文中,我们建立在以前的工作,解决定量和定性快速评估模型的困难。我们提出了DIME(数据集,索引,模型,嵌入),一个模态无关的工具,处理多模态数据集,训练模型和数据预处理器,以支持直接的模型比较与Web浏览器图形用户界面。DIME本质上支持构建与模态无关的可查询索引和提取相关特征嵌入,因此有效地兼作探索和搜索数据集的高效跨模态工具。

英文摘要:

Cross-modal retrieval relies on accurate models to retrieve relevant results for queries across modalities such as image, text, and video. In this paper, we build upon previous work by tackling the difficulty of evaluating models both quantitatively and qualitatively quickly. We present DIME (Dataset, Index, Model, Embedding), a modality-agnostic tool that handles multimodal datasets, trained models, and data preprocessors to support straightforward model comparison with a web browser graphical user interface. DIME inherently supports building modality-agnostic queryable indexes and extraction of relevant feature embeddings, and thus effectively doubles as an efficient cross-modal tool to explore and search through datasets.

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。

↑