Abstract: Filter banks are a class of signal processing techniques that can be used to reveal the local energy of a signal at multiple scales. Utilizing such filtering allows us to consider local texture and other data characteristics, and permits volume classification and visualization that cannot be accomplished easily using conventional, transfer function-based methods. Our filter bank approach increases the dimensionality, and thus, the complexity of the classification task. We have therefore developed an interactive user interface for specifying and visualizing these higher dimensional classifiers, which enables volume data exploration and visualization in a filter-bank space. We demonstrate that this technique is particularly effective for the classification of noisy data, and for classifying regions that are difficult to approach using conventional methods.