1 algoritm (GLCIA) : a superior computational method to rapid determine co-occurrence probability texture features Author : David A. Clausi, Yongping Zhao Source : Computers & Geosciences 29 (2003) 837-850 Speaker : Kai-Hung Chen Date : Dec. 8, 2004
Mar 19, 2016
1
Grey level co-occurrence integrated algoritm (GLCIA) :
a superior computational method to rapid determine
co-occurrence probability texture features
Author : David A. Clausi, Yongping Zhao
Source : Computers & Geosciences 29 (2003) 837-850
Speaker : Kai-Hung Chen Date : Dec. 8, 2004
2
Outline
Introduction
Method
Experiments
Conclusions
3
Introduction
GLCM
GLCIA GLCHS GLCHH
GLCHSH GLCHDH
Statics for co-occurrence probability
4
G=4 (0-3) θ=0 and 180 D=120 possible co-occurrence pairs
GLCM(Gray level co-occurrence matrix)
Ex:(3.2) appears 2 times in the matrix
Probability:0.1
5
Statics for co-occurrence probability
6
Method GLCHS
(Gray level co-occurrence hybrid structure)
7
Method GLCHH(Gray level co-occurrence hybrid histogram)
A:normalized sum histogram
B:normailized difference histogram
8
Method GLCHSH(Gray level co-occurrence hybrid sum histogram)
9
Method GLCHDH(Gray level co-occurrence hybrid difference histogram)
10
Experiment 1/5
Computational time:μs/sample8 statics:DIS,CON,IDM,INV,COR,UNI,ENT,MAX5 statics: CON,IDM,INV,DIS,COR4 statics: CON,IDM,INV,DIS
11
Experiment 2/5
8 statics 5 statics 4 statics
12
Experiment 3/5
13
Experiment 4/5
14
Experiment 5/5
image size:1000*1000θ=0,90,180 and 270Processor:2.0 GHz
15
Conclusions
Quickly calculate co-occurrence probability
Especially for large-scale remote-sensing image