سال انتشار: ۱۳۹۰

محل انتشار: هفتمین کنفرانس ماشین بینایی و پردازش تصویر

تعداد صفحات: ۵

نویسنده(ها):

Tahereh Hassanzadeh – Faculty of IT and Computer EngineeringQazvin Azad University, Qazvin, IRAN
Hakimeh Vojodi –
Amir Masoud Eftekhari Moghadam –

چکیده:

Multilevel thresholding is an important techniquefor image processing. Many thresholding techniques havebeen proposed in the literature. Among them, the maximumentropy thresholding (MET) has been widely applied. In thispaper is presented a novel optimal multilevel thresholdingalgorithm based on maximum entropy measure and L´evyflightFirefly algorithm (LFA) for image segmentation. Thisnew method called, the maximum entropy based on L´evyflightFirefly algorithm for multilevel thresholding(MELFAMT) method. The proposed segmentation method isemployed for five benchmark images and the performancesobtained outperform results obtained with well-knownmethods, like Gaussian smoothing method, Symmetry-dualitymethod, improved GA-based algorithm, the hybridcooperative-comprehensive learning based PSO algorithm(HCOCLPSO)and A new social and momentum componentadaptive PSO algorithm (SMCAPSO) for image segmentation.