Ranked representation of vector fields

Abstract Browsing and visualizing large datasets is often a tedious chore. Locating features, especially in a wavelet transform domain is usually offered as a possible solution. Wavelet transforms decorrelate data and facilitate progressive access through streaming. The work reported here describes a scheme that allows the user to first visualize regions containing significant features. Various region and coefficient ranking strategies can be incorporated into this approach so that a progressively encoded bitstream can be constructed. We examine four wavelet ranking schemes and demonstrate the usefulness of the feature-based schemes for a 2D oceanographic dataset.

• Type: Conference
• Year: 2003
• Where: DAGSTUHL
• Authors: Bala Krishna Nakshatrala, David Thompson, Raghu Machiraju
• Keywords: DAGSTUHL 2003, DAGSTUHL 2007, Wavelet Ranking Schemes, Wavelet Transform Domain, Wavelet Transforms