Typhoon is a severe meteorological phenomenon that heavily affects many regions in Japan. Hence the accurate analysis and prompt prediction of typhoon is an important problem to solve. However, the analysis of typhoon involves the problem of time-series analysis of typhoon cloud patterns, which challenge, the author conjectures, cannot be solved solely using tools developed in the meteorology community. In this paper, therefore, the author builds up the collection of typhoon images, and use it as a testbed for applying various techniques developed in the pattern recognition and computer vision community, and developing a new type of method for typhoon analysis. The author also refers to the result of similarity-based typhoon cloud pattern retrieval using the graph structure.

      @InProceedings{ k:cvim00,
      author = {Asanobu KITAMOTO},
      title = {'Digital Typhoon' --  Typhoon Analysis based on Artificial Intelligence Approach},
      booktitle = {Technical Report of Information Processing Society of Japan (IPSJ)},
      volume = {CVIM123-8},
      pages = {59-66},
      year = 2000,
      month = 9,
      doi = {},
      note = { (in Japanese)},
      }