Foundations and Trends in Information Retrieval
Narrative theme navigation for sitcoms supported by fan-generated scripts
Multimedia Tools and Applications
| Hi-index | 0.00 |
The content track in ACM Multimedia 2007 will for the first time, feature close to 50% of papers related to TRECVID [1]. There are more TRECVID papers in the other tracks too. A large number of accepted papers also focus on various enhanced machine learning (ML) algorithms. While TRECVID provides a large scale structured dataset for systematic research, and machine learning framework offers a principled way to handle large scale unstructured multimodal data, it is generally felt that there should be more beyond TRECVID and ML. For example, the conference seems to have little representation in a number of key emerging areas like web search, (harnessing) social intelligence, personal media management, mobile and Internet applications, multimedia advertising, and human computer interactions. Part of the positive development of TRECVID is that we now have reasonably realistic large-scale datasets from which to empirically verify many of our algorithms/techniques. However, it offers only one facet of multimedia application and there are many interesting areas and issues that should also be considered. Moreover, reviewers are now looking for complete papers that have large scale testing. This means that areas that do not have large datasets will likely not be acceptable in the conference. A related problem is that many new idea papers, that tend to address new areas with non-formal datasets, will suffer the same fate. This panel will address these issues. In particular, the panel will discuss: • Are we having too many TRECVID type papers, at the expense of other emerging areas? • What new emerging areas of multimedia research we should be targeting?" • How to encourage papers in these emerging areas of research and applications? • How to leverage users in the loop, through social networking and user generated contents etc, to setup large test sets in important emerging areas? • How to ensure that more new idea papers are presented at future conferences.