Texture synthesis via a noncausal nonparametric multiscale Markov random field

被引：132

作者：

Paget, R ^{[1
]}

Longstaff, ID

机构：

[1] Univ Queensland, Dept Elect & Comp Engn, Brisbane, Qld 4072, Australia

[2] Univ Queensland, Cooperat Res Ctr Sensor Signal & Informat Proc, Brisbane, Qld 4072, Australia

来源：

IEEE TRANSACTIONS ON IMAGE PROCESSING | 1998年 / 7卷 / 06期

关键词：

local annealing; Markov random fields; multiresolution; nonparametric estimation; texture synthesis;

D O I：

10.1109/83.679446

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Our noncausal, nonparametric, multiscale, Markov random field (MRF) model is capable of synthesizing and capturing the characteristics of a wide variety of textures, from the highly structured to the stochastic. We use a multiscale synthesis algorithm incorporating local annealing to obtain larger realizations of texture visually indistinguishable from the training texture.

引用

页码：925 / 931

页数：7

共 17 条

[1] BESAG J, 1974, J ROY STAT SOC B MET, V36, P192
[2] BESAG J, 1986, J R STAT SOC B, V48, P259
[3] MULTIPLE RESOLUTION SEGMENTATION OF TEXTURED IMAGES
BOUMAN, C
LIU, BD
[J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1991, 13 (02) : 99 - 113
[4] Brodatz P, 1966, TEXTURES PHOTOGRAPHI
[5] A PARALLEL IMAGE SEGMENTATION ALGORITHM USING RELAXATION WITH VARYING NEIGHBORHOODS AND ITS MAPPING TO ARRAY PROCESSORS
DERIN, H
WON, CS
[J]. COMPUTER VISION GRAPHICS AND IMAGE PROCESSING, 1987, 40 (01): : 54 - 78
[6] GEMAN D, 1990, LECT NOTES MATH, V1427, P113
[7] STOCHASTIC RELAXATION, GIBBS DISTRIBUTIONS, AND THE BAYESIAN RESTORATION OF IMAGES
GEMAN, S
GEMAN, D
[J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1984, 6 (06) : 721 - 741
[8] A RENORMALIZATION-GROUP APPROACH TO IMAGE-PROCESSING PROBLEMS
GIDAS, B
[J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1989, 11 (02) : 164 - 180
[9] Haindl M., 1991, CWI Quarterly, V4, P305
[10] Heeger D. J., 1995, Proceedings. International Conference on Image Processing (Cat. No.95CB35819), P648, DOI 10.1109/ICIP.1995.537718

← 1 2 →