IIM Future and Opportunities

Abstract

In recent years, our ability to acquire and store large quantities of data has greatly surpassed our ability to access and meaningfully extract information from it. For example, EO archives store many millions of images, but only a small percentage of this data is used in applications. Meanwhile, state of the art methods for data or phenomena understanding of relevance for science or engineering are no longer simply based on theory followed by experiments. Sophisticated computations can complement or substitute experiments enabling, exploration, understanding or control of highly complex systems. In addition, investigations of natural phenomena in more flexible, faster or cheaper way are supported. However, the current methods are limited in several important aspects: - they are not effective for analyzing distributed heterogeneous data sets; - they do not identify cause-effected relationships; - they not enable a close man-machine collaboration; - many of the information systems follow a straight forward strategy not being able to learn; - scientific and engineering communities need much higher computing performance. The article presents an overview of the state of the art in Image Information mining (IIM), Knowledge Discovery and Data Mining (KDD), and related methods for accessing and analysing EO data including a discussion of new theoretical developments in fields like: - theory of information transmission and semantic coding - modelling of human conjecture, and semiotic aspects - information representation for communication and understanding - data and information visualization, grammars of the visual language - multimodal data and information fusion, and augmentation of the data with meaning - knowledge organization and dissemination methods for education and outreach - human-centered internet tools, like advanced browsers, filters, visual navigators A proposal for innovative research activities is presented aiming at enhanced exploitation of EO data.