Generic Content-based News Picture Retrieval with Local Invariant Features

图像检索从大型数据库中检索视觉内容满足用户提交要求的图像。除了直接使伙伴组织受益外，该项目还将使新闻机构和出版商、数字图书馆和电影档案馆能够更有效地访问大型图片库。它将使公众使用视觉信息更加方便和经济。它将帮助保安人员从大型闭路电视网络生成的图像中轻松准确地找到特定场景。此外，开发的技术可以应用于远程教育和电子商务。在这个项目中开发的新算法将使澳大利亚和世界科学界受益。

Image Retrieval searches for images from large databases whose visual content meets the requirements submitted by users. Besides directly benefiting the Partner Organization, this project will enable more efficient access to large picture repositories in news agencies and publishers, digital libraries and film archives. It will make public use of visual information much more convenient and economical. It will help security officers to effortlessly and accurately find particular scenes from the images generated by a large closed-circuit TV networks. Also, the developed technology can be applied to tele-education and e-commerce. New algorithms developed in this project will benefit the Australian and world scientific communities.