000			02201nam a22002777a 4500
003			KOHA_Geminibilgi
005			20220429154417.0
008			220429d2022 cy ||||| m||| 00| 0 eng d
040			_aCY-NiCIU _beng _cCY-NiCIU _erda
041			_aeng
090			_aD 304 _bY47 2022
100	1		_aYırtıcı, Tolga
245	1	0	_aIMPROVED TURKISH SIGN LANGUAGE RECOGNITION USING CONVOLUTIONAL NEURAL NETWORKS / _cTOLGA YIRTICI; SUPERVISOR: ASST. PROF. DR. KAMİL YURTKAN
264			_c2022
300			_a114 sheets; _c31 cm. _eIncludes CD
336			_2rdacontent _atext _btxt
337			_2rdamedia _aunmediated _bn
338			_2rdacarrier _avolume _bnc
502			_aThesis (PHD) - Cyprus International University. Institute of Graduate Studies and Research Computer Engineering Department
504			_aIncludes bibliography (sheets 106-114)
520			_aABSTRACT This thesis started with idea of creating a robust system for Turkish Sign Language recognition. Fingerspelling of Turkish Sign Language alphabet is chosen for these purposes. Turkish Sign Language alphabet consists of 29 letters just like in speaking language. Alphabet letters can be used to form a word. Two different systems designed in this manner, one with a region-based object detection method and other is an information content-based feature selection. The designed systems are employed with AlexNet architecture using transfer learning. AlexNet is a pre-trained Convolutional Neural Network that utilized for classification problems. The novel object detection method is tested with three different algorithms and achieved the best result of 0.997 mean Average Precision and 0.9982 accuracy rate. The information content-based feature selection method with employed the same AlexNet architecture, used a novel feature selection algorithm and achieved more than 80% accuracy rate. Both of the systems are trained and tested on the dataset created for this study in a studio.
650		0	_aObject-oriented methods (Computer science) _vDissertations, Academic
650		0	_aTransfer learning (Machine learning) _vDissertations, Academic
700	1		_aYurtkan, Kamil _esupervisor
942			_2ddc _cTS
999			_c284277 _d284277