Threshold Modification Model for the Optimal Evolution Algorithm
2012 ◽
Vol 239-240
◽
pp. 1104-1107
Keyword(s):
The number of pixels of object and background generally has a large difference. Thresholds calculated by the optimal evolution algorithm (OEA) tend to the class which has more samples than the other, then the optimal evolution algorithm falls into a local optimal solution. A threshold modification model was built up for an improved optimal evolution algorithm. The improved OEA can adaptively modify the threshold in each evolutional generation according to the difference of pixels’ number between object and background. Experimental results show that the improved OEA is more stable, converge more rapidly, and has better segmentation quality than OEA.
2013 ◽
Vol 756-759
◽
pp. 3231-3235
2013 ◽
Vol 380-384
◽
pp. 3854-3857
2010 ◽
Vol 143-144
◽
pp. 1204-1206
2012 ◽
Vol 433-440
◽
pp. 5596-5600
2012 ◽
Vol 452-453
◽
pp. 1491-1495
2013 ◽
Vol 300-301
◽
pp. 146-149
◽
2015 ◽
Vol 713-715
◽
pp. 1583-1588
Keyword(s):
1926 ◽
Vol 99
(697)
◽
pp. 326-339
◽
Keyword(s):
1973 ◽
Vol 29
(02)
◽
pp. 490-498
◽