6. Conclusion

This chapter describes the creation of video summaries and visualization systems through multimodal feature analysis. The integration of multiple features provides insights into video characterization not present with a single feature. Multimodal analysis is necessary for combining multiple forms of image, audio and language information, as well as incorporating new forms of metadata in the future. Surrogates have proven advantages for allowing information from a video document to be found and accessed quickly and accurately. Visualization techniques addressing text corpora and databases have been shown to be applicable to video libraries as well.

Access to video information is greatly improved through advances in video summarization and visualization. The video summary provides an efficient means for abstracting, shortening or simply browsing a large collection. Future digital video library interfaces that summarize sets of video documents and leverage from the library's multiple media and rich visual nature need to be designed and evaluated so that the wealth of material within such libraries can be better understood and efficiently accessed.