Visualizing vector fields is one of the most important and attractive research areas in scientific visualization. Several techniques are known in the literature; some traditional approaches use 2D/3D arrows or particle traces, while other  methodologies display vector fields by dense or sparse textures. This paper focuses on the clustering problem arising for dense texture algorithms where some low contrast areas can appear in images thus reducing the capability of investigating flow details. The proposed method assigns pixel colors according to a local contrast analysis phase. In this way streamlines denoting flow characteristics are always well distinguishable.

keywords: vector field visualization, scientific visualization, texture based algorithms, LIC, local contrast.