ECONOMETRICS Econometrics

Econometrics

1. Introduction

Two parallel analyses are carried out in order to assess the degree of integration of financial markets within Europe, within East Asia, between these two regions, and with the external financial community. The investigation is based on cointegration and Granger causality techniques, to detect the presence of short-run and long-run cross-country relationships in equity and bond markets. The empirical analysis performed for seven European and eleven East Asian financial markets confirms that in Europe financial integration is significantly more advanced than in East Asia. It provides evidence in favour of the fact that the level of integration between bond markets is higher than between equity markets within Europe, whilst the opposite holds true in the East Asian region. Finally, East Asian bond markets display little evidence of co-movement, despite the recent initiatives launched in order to increase financial integration in the region. These techniques generally fall into one of the following research areas.

• Video indexing: research in this area aims at creating compact indices for large video databases and providing easy browsing and intelligent query mechanisms (Ardizzone et al., 1996; Chen et al., 1998; Kobla et al., 1996; Naphade et al., 1998). Potential applications include multimedia databases, digital libraries, and web media portals.

• Video filtering and abstraction: research in this area tries to generate an abstract version of the video content that is important or interesting by extracting key portions of the video (Saur et al., 1997). This can be used for personalized video delivery, intelligent digital video recording devices, and video summaries for large multimedia archives.

• Audio indexing and analysis: research in this relatively new area includes audio classification, audio indexing and retrieval, music retrieval, etc (Patel and Sethi, 1995; Sundaram and Chang, 2000a; Zhang and Kuo, 1999b). Efforts have also been made in combining audio information with visual information to help index and analyze video content.

Besides these general areas, there is also some research with more specific objectives, which exploits knowledge in various domains. This has produced some interesting applications such as event detection in sports programs, anchorperson detection in news, and so on. However, techniques proposed in these applications are often limited to their application domains.

In all these areas, one critical and challenging issue is to extract features from multimedia data. Low-level features are compact, mathematical representations of the physical properties of the video and audio data. They greatly reduce the amount of data to be analyzed, provide metrics for comparison, and serve as the foundation for indexing, analysis, high-level understanding, and classification. Much research work has been done to investigate various visual and audio features, their extraction methods, and their application in various domains. In fact, the ISO/IEC International Standard MPEG-7 has specified a set of descriptors (i.e., features) and description schemes for the description of multimedia content.

In this paper we will give an overview of state-of-the-art algorithms and tools using video and audio features. In particular, we will investigate features that can be efficiently extracted from compressed video and audio ...