2600

2635پ

2759

كارشناسي ارشد

مخابرات

اصفهان: دانشگاه صنعتي اصفهان، دانشكده برق و كامپيوتر

1383

نه، 74، [II]ص .: مصور، جدول، شكل، عكس، نمودار

1395/11/20

كتابنامه

برق و كامپيوتر

مهندسي برق و كامپيوتر

ID2759

به فارسي و انگليسي : قابل رويت درنسخه ديجيتالي

Abstract Over the last thirty years researchers in neural networks image processing andcomputer vision have investigated a number of issues related to face recognition by humanand machine Machine recognitIon of faces from still and video images has severalapplications such as static matching of controlled format photographs such as passports credit cards drivers license and mug shots to real time matching of surveillance videoimages Such applications have different constraints in terms of complexity of processingrequirements and consclusively a wide different technical challenges are presented Existing techniques and systemshave beenexamined on different sets of images of variouscomplexity Automatic face recognition is one of the difficult challenges in patternrecognition Usually human faces are similar in structure and very little differences indifferent faces Illumination conditions facial expression and face orientation result in thatface recognition be one of the most difficult problem in pattern recognition At the beginning of this thesis several methods of face recognition are reviewd Thenone of the statistical robust methods with Hidden Markov Model is studied and result ofsimulation of face recogniton by continous Hidden Markov Model using Karhunen Loeveexpansion with Matlab software is presented And then face recogniton simulated by MLPand RBF neural networks and these results are compared At the end of thesis structure ofRBF compared with HMM and summary of results are presented r It I d v J I

سعيد صدري

علي محمد دوست حسيني

محمود مدرس هاشمي، رسول امير فتاحي