#FIG 3.2LandscapeCenterInchesLetter  100.00Single-21200 21 1 0 2 0 7 50 0 -1 0.000 1 0.0000 3450 4500 300 150 3450 4500 3750 45002 1 0 1 0 7 50 0 -1 0.000 0 0 -1 1 0 2	1 1 1.00 60.00 120.00	 4200 1800 4800 18002 1 0 1 0 7 50 0 -1 0.000 0 0 -1 1 0 2	1 1 1.00 60.00 120.00	 5550 2250 5550 28502 1 0 1 0 7 50 0 -1 0.000 0 0 -1 1 0 2	1 1 1.00 60.00 120.00	 4800 3300 4200 33002 1 0 1 0 7 50 0 -1 0.000 0 0 -1 1 0 2	1 1 1.00 60.00 120.00	 3450 2850 3450 22502 1 0 1 0 7 50 0 -1 0.000 0 0 -1 1 0 2	1 1 1.00 60.00 120.00	 3450 3750 3450 43502 2 0 2 0 7 50 0 -1 0.000 0 0 -1 0 0 5	 2700 2850 4200 2850 4200 3750 2700 3750 2700 28502 2 0 2 0 7 50 0 -1 0.000 0 0 -1 0 0 5	 4800 2850 6300 2850 6300 3750 4800 3750 4800 28502 2 0 2 0 7 50 0 -1 0.000 0 0 -1 0 0 5	 4800 1350 6300 1350 6300 2250 4800 2250 4800 13502 2 0 2 0 7 50 0 -1 0.000 0 0 -1 0 0 5	 2700 1350 4200 1350 4200 2250 2700 2250 2700 13502 1 0 1 0 7 50 0 -1 0.000 0 0 -1 1 0 2	1 1 1.00 60.00 120.00	 2100 2100 2700 21002 1 0 1 0 7 50 0 -1 0.000 0 0 -1 1 0 2	1 1 1.00 60.00 120.00	 2100 1500 2700 15004 0 0 50 0 16 10 0.0000 4 150 1215 2775 1950 pixel using cluster\0014 0 0 50 0 16 10 0.0000 4 150 765 2775 2145 parameters\0014 0 0 50 0 16 10 0.0000 4 120 900 2775 1575 Perform soft\0014 0 0 50 0 16 10 0.0000 4 120 1425 2775 1770 classification of each\0014 0 0 50 0 16 10 0.0000 4 150 1290 2775 3075 Measure goodness\0014 0 0 50 0 16 10 0.0000 4 150 1410 2775 3270 of fit using Rissanen\0014 0 0 50 0 16 10 0.0000 4 150 1380 2775 3660 far,then save result.\0014 0 0 50 0 16 10 0.0000 4 120 1350 4875 1575 Re-estimate cluster\0014 0 0 50 0 16 10 0.0000 4 150 1410 4875 2160 covariance matrices)\0014 0 0 50 0 16 10 0.0000 4 150 810 4875 1770 parameters.\0014 0 0 50 0 16 10 0.0000 4 120 1245 4875 3075 Reduce number of\0014 0 0 50 0 16 10 0.0000 4 150 1200 3600 4125 Only one cluster?\0014 0 0 50 0 16 10 0.0000 4 150 1290 4875 1965 (mean vectors and\0014 0 0 50 0 16 10 0.0000 4 120 1515 2775 3465 criterion. If best fit so\0014 0 0 50 0 16 10 0.0000 4 150 735 4875 3270 clusters by\0014 0 0 50 0 16 10 0.0000 4 150 1020 4875 3465 combining two\0014 0 0 50 0 16 10 0.0000 4 150 765 975 2175 parameters\0014 0 0 50 0 16 10 0.0000 4 120 1035 975 2025 Initialize cluster\0014 0 0 50 0 16 10 0.0000 4 120 720 975 1650 of clusters\0014 0 0 50 0 16 10 0.0000 4 120 1095 975 1500 Initialize number\0014 1 0 50 0 16 10 0.0000 4 120 285 3427 4560 Exit\0014 0 0 50 0 16 10 0.0000 4 120 1125 4875 3653 nearest clusters.\001