A new framework for characterization of halftone textures

Characterization of halftone textures is important for quantitative assessment of halftone quality. We develop a new framework based on directional local sequency analysis and a filter bank structure. We decompose a halftone image into subband images from which we can easily reconstruct the original halftone. Based on these subband images, we define the directional sequency spectrum which is analogous to the 2D Fourier spectrum, and formulate several texture measures. Three test images are used to justify these measures.