IDLP: International Collaboration to Advance User-oriented Technologies for Managing and Distributing Images in Digital Libraries

该项目将开发面向用户的数字图书馆分发技术的形象管理。一个由美国、中国和台湾的计算机和信息科学家组成的跨学科团队将研究使用集成方法搜索数字图像集合的有效方法。美国的研究团队由西蒙斯学院、宾夕法尼亚大学和宾夕法尼亚州立大学的教职员工和学生组成。将研究几个技术领域：(1)基于本体的图像检索；(2)基于机器学习的图像检索，包括基于对象的局部图像搜索；(3)图像自动标注；(4)知识产权保护技术。该团队将使用真实世界的数字图书馆数据集来开发适合实际部署的面向用户的技术。值得注意的是，这项研究将利用现有的大量与中国皇帝有关的各种分辨率和巨大文化意义的图像以及现有丰富的描述性信息。该项目将促进对数字图像管理和分发技术的理解。这可能会导致从大型集合中检索图像的更高效和有效的方法。它将通过提供自动图像注释来提供减少劳动密集型活动的手段。它将利用该项目将要开发的访问模型和知识产权技术，使分布在不同地点的高质量图像收集能够更有意愿地联网。该项目对于全面解决面向用户的图像相关技术问题具有重要意义。由此产生的技术可以应用于其他学科内容和技术驱动的领域，如计算机视觉。在积极和更广泛的影响中，将加强数字图书馆研究的国际基础设施，提供比目前可能的更大规模的普遍获取文化和遗产的机会，以及扩大全球数字图书馆研究界的参与。该项目还将通过培训跨学科研究生和为本科生提供新的跨学科课程来满足国家对科学家的需求。这一国际伙伴关系将促进小型学院和少数族裔学生参与这类先进的多学科研究。

This project will develop user-oriented image management of distribution technologies for digital libraries. An interdisciplinary team of computer and information scientists from US, China, and Taiwan will investigate efficient ways to search digital collections of images using an integrated approach. The US team of researchers consists of faculty members and students of Simmons College, University of Pennsylvania, and The Pennsylvania State University. Several technology areas will be studied: (1) ontology-based image retrieval, (2) machine learning based image retrieval including object-based partial image search, (3) automatic image annotation, and (4) intellectual property (IP) protection techniques. The team will use real-world digital library datasets to develop user-oriented technologies suitable for practical deployment. Notably, the research will utilize an existing collection consisting of a large quantity of images associated with the first Emperor of China of all types of resolution and with enormous cultural significance as well as the existing rich descriptive information.This project will advance understanding of digital imagery management and distribution technologies. It is likely to result in more efficient and effective methods for retrieving images from large collections. It will provide means to decrease labor-intensive activities by providing automatic image annotations. It will enable more willing networking of quality image collections in distributed locations with the access model and IP techniques to be developed by the project. The project has significance for addressing the issue of user-oriented image-related technologies in a general way. The resulting techniques can be applied to other subject contents as well as technology-driven domains, such as computer vision.Among the positive broader impacts will be enhancement of the international infrastructure for digital library research, providing larger-scale universal access to culture and heritages than what is possible now, and expanding the participation of the global digital library research community. This project will also address the national need for scientists by training interdisciplinary graduate students and by offering new interdisciplinary courses for undergraduate students. This international partnership will promote the involvement of small college and minority students in this type of advanced multidisciplinary research.