Explicit Frequency Control for High-Quality Texture-Based Flow Visualization

In this work we propose an effective method for frequency-controlled dense flow visualization derived from a generalization of the Line Integral Convolution (LIC) technique. Our approach consists in considering the spectral properties of the dense flow visualization process as an integral operator defined in a local curvilinear coordinate system aligned with the flow. Exploring LIC from this point of view, we suggest a systematic way to design a flow visualization process with particular local spatial frequency properties of the resulting image. Our method is efficient, intuitive, and based on a long-standing model developed as a result of numerous perception studies. The method can be described as an iterative application of line integral convolution, followed by a one-dimensional Gabor filtering orthogonal to the flow. To demonstrate the utility of the technique, we generated novel adaptive multi-frequency flow visualizations, that according to our evaluation, feature a higher level of frequency control and higher quality scores than traditional approaches in texture-based flow visualization.