MODIFIED WIGNER DISTRIBUTION AND APPLICATION TO WRITER IDENTIFICATION
Liu Cheng-li · Chinese Journal of Computers · 1997
In this paper, a simplified Wigner distribution is proposed and applied to writer identification (WI). Wigner distribution represents the local spectrum of image but is expensive in computation and storage. It is proved that Wigner distribution is redundant and then its representation is simplified. The simplified Wigner distribution is information-preserving and still powerful in texture measuring. Experiments of WI show that this method is better than previous handwriting textural methods.