Arding Hsu

Content-based browsing of video sequences

A novel methodology to represent the contents of a video sequence is presented. The representation is used to allow the user to rapidly view a video sequence in order to find a particular point within the sequence and/or to decide whether the contents of the sequence are relevant to his or her needs. This system, referred to as content-based browsing, forms an abstraction to represent each shot of the sequence by using a representative frame, or an Rframe, and it includes management techniques to allow the user to easily navigate the Rframes. This methodology is superior to the current techniques of fast forward and rewind because rather than using every frame to view and judge the contents, only a few abstractions are used. Therefore, the need to retrieve the video from a storage system and to transmit every frame over the network in its entirety no longer exists, saving time, expenses, and bandwidth.

Image processing on encoded video sequences

This paper presents a novel approach to processing encoded video sequences prior to complete decoding. Scene changes are easily detected using DCT coefficients in JPEG and MPEG encoded video sequences. In addition, by analyzing the DCT coefficients, regions of interest may be isolated prior to decompression, increasing the efficiency of any subsequent image processing steps, such as edge detection. The results are currently used in a video browser and are part of an ongoing research project in creating large video databases. The procedure is detailed with several examples presented and studied in depth.

Medical image retrieval by spatial features

A content-based indexing technique is proposed. The image features used are derived from the relative spatial relationships among internal image entities. The similarity measurement is based on causality (probability) which indicates the degree of similarity between a users query and images. The index structure contains a set of causality-based similarity trees with nodes connected to an information causal net. For a given (weighted) query, the initial set of similar images is identified via similarity trees and then refined through the information causal net. The method is introduced with an example using magnetic resonance chest images.<<ETX>>

The management and applications of teleaction objects

Teleaction objects (TAOs) possess private knowledge specific to the object instances. The user can create and modify the private knowledge of a TAO, so that it automatically reacts to certain events to preperform operations for generating timely responses, improving operational efficiency and maintaining consistency. Moreover, TAOs also possess a hypergraph structure leading to the effective presentation and efficient communication of multimedia information. The active multimedia system (AMS) is designed to manage the TAOs. In the AMS, each TAO has its own private knowledge, realized by the active index cells. The applications to smart multimedia mail and multimedia information retrieval are described to illustrate the usefulness of TAOs. We discuss the advantages of the proposed approach and future research.

Efficient and reliable digital media archive for content-based retrieval

Abstract. Recent years have witnessed a significant price reduction in many enabling technologies for wide-spread deployment of multimedia to desktop PCs and workstations. This advancement has lead to an increasing demand for systems that can store, retrieve, and manipulate large volumes of multimedia information. For a multimedia information system to better meet information users needs, it must provide suitable access structures and methods. The answers to this demand fall into the research area of what most people called content-based multimedia indexing and retrieval. Existing approaches to content-based indexing and retrieval have limitations. What we need is a digital-media-archiving system that is both efficient and reliable. By reliable, we mean that users should be able to retrieve documents that have the most potential for being relevant to their queries. On the other hand, an efficient digital-media-archiving system should provide an environment that allows human operators to create document indices without the need to manually watch every multimedia object and enter keyword descriptions. This can be done by providing initial structure information of the video to the user, by guiding the human operator through the indexing process, and by offering tools to create multiple media representations in a hierarchical structure.