Abstract. In remote sensing hyperspectral image processing, identifying the constituent spectra (endmembers) of the materials in the image is a key procedure for further analysis. ...
This paper introduces a uniform statistical framework for both 3-D and 2-D object recognition using intensity images as input data. The theoretical part provides a mathematical too...
There have been important recent advances in object recognition through the matching of invariant local image features. However, the existing approaches are based on matching to i...
Automatic image tagging is important yet challenging due to the semantic gap and the lack of learning examples to model a tag’s visual diversity. Meanwhile, social user tagging ...
Communication across the science-policy interface is complicated by uncertainty and ignorance associated with predictions on which to base policies. The international symposium â€...