Statistical Pattern Spectrum for Binary Pattern Recognition

A new shape descriptor based on Statistical Morphology is presented. The descriptor, called specstrum, is a statistical extension of pattern spectrum, (i.e., pecstrum) useful to represent shape information related to binary patterns in noisy conditions. The major features of the specstrum are an increased stability under varying noisy conditions and a more regular shape description capability. Results are presented for a parking surveillance application.

[1]  C. D. Gelatt,et al.  Optimization by Simulated Annealing , 1983, Science.

[2]  J. Goutsias,et al.  Optimal Morphological Pattern Restoration from Noisy Binary Images , 1991, IEEE Trans. Pattern Anal. Mach. Intell..

[3]  Ioannis Pitas,et al.  Nonlinear Digital Filters - Principles and Applications , 1990, The Springer International Series in Engineering and Computer Science.

[4]  Donald Geman,et al.  Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[5]  Gian Luca Foresti,et al.  Statistical morphological filters for binary image processing , 1994, Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing.

[6]  Jean Serra,et al.  Image Analysis and Mathematical Morphology , 1983 .

[7]  A. Venetsanopoulos,et al.  The classification properties of the pecstrum and its use for pattern identification , 1991 .

[8]  Petros Maragos,et al.  Morphological filters-Part II: Their relations to median, order-statistic, and stack filters , 1987, IEEE Trans. Acoust. Speech Signal Process..

[9]  Anastasios N. Venetsanopoulos,et al.  Multidimensional shape description and recognition using mathematical morphology , 1988, J. Intell. Robotic Syst..