OPTIMIZATION OF THE WEIGHTED MEDIAN FILTER BY LEARNING

被引:18
作者
ASANO, A
ITOH, K
ICHIOKA, Y
机构
[1] Department of Applied Physics, Osaka University, Suita, Osaka, 565
关键词
D O I
10.1364/OL.16.000168
中图分类号
O43 [光学];
学科分类号
070207 ; 0803 ;
摘要
The weighted median filter (WMF) is a generalization of the median filter. The WMF is more effective for image processing than the conventional median filter. However, the design of the parameters of the WMF is a difficult problem. A novel method of optimizing the WMF is proposed that utilizes the close relation between the nonrecursive WMF and the feed-forward neural network with shift-invariant weight coefficients. The optimization problem of the WMF results in the learning of the interconnection weights of the network.
引用
收藏
页码:168 / 170
页数:3
相关论文
共 6 条
[1]   CONVERGENCE PROPERTIES OF RECURSIVE RANK-ORDER FILTER AND NEURAL NETWORK [J].
ASANO, A ;
ZHANG, W ;
ITOH, K ;
ICHIOKA, Y .
PATTERN RECOGNITION LETTERS, 1990, 11 (08) :557-560
[2]   BIPOLAR MORPHOLOGY AND ITS APPLICATIONS [J].
ASANO, A ;
ITOH, K ;
ICHIOKA, Y .
JAPANESE JOURNAL OF APPLIED PHYSICS PART 2-LETTERS & EXPRESS LETTERS, 1990, 29 (07) :L1270-L1273
[3]   THE WEIGHTED MEDIAN FILTER [J].
BROWNRIGG, DRK .
COMMUNICATIONS OF THE ACM, 1984, 27 (08) :807-818
[4]  
Rumelhart D.E., 1986, PARALLEL DISTRIBUTED, V1, P318
[5]   STACK FILTERS [J].
WENDT, PD ;
COYLE, EJ ;
GALLAGHER, NC .
IEEE TRANSACTIONS ON ACOUSTICS SPEECH AND SIGNAL PROCESSING, 1986, 34 (04) :898-911
[6]   PARALLEL DISTRIBUTED-PROCESSING MODEL WITH LOCAL SPACE-INVARIANT INTERCONNECTIONS AND ITS OPTICAL ARCHITECTURE [J].
ZHANG, W ;
ITOH, K ;
TANIDA, J ;
ICHIOKA, Y .
APPLIED OPTICS, 1990, 29 (32) :4790-4797