STIMULATE: Multimodal Indexing, Retrieval, and Browsing: Combining Content-Based Image Retrieval with Text Retrieval

点击翻译按钮获取中文摘要

*** In the rapidly emerging area of multimedia information systems, effective indexing and retrieval techniques are critically important. In this project, the Center for Intelligent Information Retrieval (CIIR) will study algorithms and mechanisms that facilitate the.process of indexing and retrieving images, video, and associated text. Specifically, the CIIR will: 1) Develop techniques for browsing and retrieving images and videos by content. Image indexing is an emerging technology which is poorly understood. New and powerful techniques will be developed for indexing images and videos using visual attributes such as color, texture and 'appearance'; 2) Develop techniques for indexing images and videos using text present in the images. Methods of assisting with manual annotation of images will also be studied. The text indexing methods will be combined using the INQUERY probabilistic retrieval engine; 3) Develop and evaluate interactive retrieval and browsing interfaces for image, video, and text retrieval. These interfaces will be based on novel visualization and user interface work done at CIIR. The CIIR will perform user studies to evaluate the appeal and effectiveness of the interfaces. In summary, the project will integrate innovative image matching, sophisticated text retrieval and new query formulation and visualization methods to create a powerful platform for indexing and retrieving images, video, and text.***