DSpace Arşivi :: by Yazar "Uçan, Osman N." değerine göre listeleniyor

Yazar "Uçan, Osman N." seçeneğine göre listele

Listeleniyor 1 - 20 / 21

Adaptive Wiener-turbo system and adaptive Wiener-turbo systems with JPEG & bit plane compressions
(2007) Buyukatak, Kenan; Uçan, Osman N.; Göse, Ersin; Osman, Onur; Kent, Sedef
In order to improve unequal error protection and compression ratio of 2-D colored images over wireless environment, we propose two new schemes denoted as ''Adaptive Wiener-Turbo System (AW-TS) and Adaptive Wiener-Turbo Systems with JPEG & Bit Plane Compressions (AW-TSwJBC)''. In AW-TS, there is a feedback link between Wiener filtering and Turbo decoder and process iteratively. The scheme employs a pixel-wise adaptive Wiener-based Turbo decoder and uses statistics (mean and standard deviation of local image) of estimated values of local neighborhood of each pixel. It has extra-ordinary satisfactory results of both bit error rate (BER) and image enhancement performance for less than 2 dB Signal-to-Noise Ratio (SNR) values, compared to separately application of traditional turbo coding scheme and 2-D filtering. In AW-TSwJBC scheme, 2-D colored image is passed through a color & bit planes� slicer block. In this block, each pixel of the input image is partitioned up to three main color planes as R,G,B and each pixel of the color planes is sliced up to N binary bit planes, which corresponds to binary representation of pixels. Thus depending on importance of information knowledge of the input image, pixels of each color plane can be represented by fewer number of bit planes. Then they are compressed by JPEG prior to turbo encoder. Hence, two consecutive compressions are achieved regarding the input image. In 2-D images, information is mainly carried by neighbors of pixels. Here, we benefit of neighborhood relation of pixels for each color plane by using a new iterative block, named as ''Adaptive Wiener-Turbo'' scheme, which employs Turbo decoder, JPEG encoder/decoders and Adaptive Wiener Filtering.
Automatic colon segmentation using cellular neural network for colorectal polyps detection
(2007) Kılıç, Niyazi; Osman, Onur; Uçan, Osman N.; Demirel, Kemal
In this paper, an automatic colon segmentation method for Computed Tomography (CT) colonography is presented. Colon segmentation is considered in order to prevent the time consumption while searching polyps out of the colon region and reduce radiologists’ interpretation time. The proposed method is the combination of pre-processing and Cellular Neural Networks (CNN). Also recurrent perceptron learning algorithm (RPLA) is used for CNN training. Original CT images are passed through a threshold and then CNN is used to erase unrelated small objects and smooth sharp corners. It is expected automatic colon segmentation will improve the radiologists’ diagnostic performance.
Blind equalization of space-time-turbo trellis coded/contilluous phase modulation over Rician fading channels
(Wiley, 2004) Uçan, Osman N.; Osman, Onur
In this paper. to improve bit error performance and bandwidth efficiency. we combine space-time blockcodes (STBC), turbo trellis codes and continuous phase modulation and denote space-time-turbo trellis coded/continuous phase modulation (ST-TTC/CPM). For high data transmission over wireless fading! channels, STBC provide the maximal possible diversity advantage for multiple decoding algorithm. We present continuous phase modulation (CPM) for ST-TTC signal. mince CPM provided low-Spectral occupancy and is suitable for power and bandwidth-limited channels. In our model. to utilize STBC efficiently. we need to estimate the channel parameters. which influence the signals having continuity property. Therefore, we develop a blind maximum likelihood channel estimation algorithm for signal propagating through a Rician fading channel. Here. Baum-Welch (BW) algorithm based on hidden Markov model (HMM) is modified to provide computationally efficient channel parameter estimation. We also investigate the performance of ST-TTC/CPM in the case of no channel state information (CSI) for various Rician parameters K and Doppler frequency. Copyright (C) 2004 AEI.
Blind equalization of turbo trellis-coded partial-response continuous-phase modulation signaling over narrow-band rician fading channels
(2005) Osman, Onur; Uçan, Osman N.
In this paper, a blind maximum-likelihood channel estimation algorithm is developed for turbo trellis-coded/continuous-phase modulation (TTC/CPM) signals propagating through additive white Gaussian noise (AWGN) and Rician fading environments. We present CPM for TTC signals, since it provides low spectral occupancy and is suitable for power- and bandwidth-limited channels. Here, the Baum-Welch (BW) algorithm is modified to estimate the channel parameters. We investigate the performance of TTC/CPM for 16-CPFSK over AWGN and Rician channels for different frame sizes, in the case of ideal channel state information (CSI), no CSI, and BW estimated CSI. © 2005 IEEE.
Breast MR segmentation and lesion detection with cellular neural networks and 3D template matching
(2008) Ertaş, Gökhan; Gülçür, H.Özcan; Osman, Onur; Uçan, Osman N.; Tunacı, Mehtap; Dursun, Memduh
A novel fully automated system is introduced to facilitate lesion detection in dynamic contrast-enhanced, magnetic resonance mammography (DCE-MRM). The system extracts breast regions from pre-contrast images using a cellular neural network, generates normalized maximum intensity-time ratio (nMITR) maps and performs 3D template matching with three layers of 12 × 12 cells to detect lesions. A breast is considered to be properly segmented when relative overlap > 0.85 and misclassification rate < 0.10. Sensitivity, false-positive rate per slice and per lesion are used to assess detection performance. The system was tested with a dataset of 2064 breast MR images (344 slices × 6 acquisitions over time) from 19 women containing 39 marked lesions. Ninety-seven percent of the breasts were segmented properly and all the lesions were detected correctly (detection sensitivity = 100 %), however, there were some false-positive detections (31%/lesion, 10%/slice). © 2007 Elsevier Ltd. All rights reserved.
Channel equalization and noise reduction based on turbo codes
(Institute of Electrical and Electronics Engineers Inc., 2003) Büyükatak, Kenan; Göse, Ersin; Uçan, Osman N.; Kent, Sedef; Osman, Onur
A new class of error control coding with binary parallel concatenated recursive systematic convolution codes turbo codes- And its an application over SAR (Synthetic Aperture Radar) image transmitting is presented in this paper. The remarkable coding gains achieved by turbo code have a great influence in various area of digital communication, including cellular mobile, deep space communications and satellite communication. The basic idea is that two or more a posteriori probability decoders exchange soft information [I]. One of the decoders calculates the a posteriori probability distribution of the information sequence and passes that information to the next decoder. The new decoder uses this information and computes its own version of the probability distribution. This exchange of information is called an iteration. After a certain number of iterations, a decision is made at the second decoder. For each iteration, the probability that we decode in favor of the correct decision will improve. At the last iteration, the hard decision is made using the soft decision of the last decoder [5]. © 2003 IEEE.
Colonic polyp detection in CT colonography with fuzzy rule based 3D template matching
(2009) Kılıç, Niyazi; Uçan, Osman N.; Osman, Onur
In this paper, we introduced a computer aided detection (CAD) system to facilitate colonic polyp detection in computer tomography (CT) data using cellular neural network, genetic algorithm and three dimensional (3D) template matching with fuzzy rule based tresholding. The CAD system extracts colon region from CT images using cellular neural network (CNN) having A, B and I templates that are optimized by genetic algorithm in order to improve the segmentation performance. Then, the system performs a 3D template matching within four layers with three different cell of 8×8, 12×12 and 20×20 to detect polyps. The CAD system is evaluated with 1043 CT colonography images from 16 patients containing 15 marked polyps. All colon regions are segmented properly. The overall sensitivity of proposed CAD system is 100% with the level of 0.53 false positives (FPs) per slice and 11.75 FPs per patient for the 8×8 cell template. For the 12×12 cell templates, detection sensitivity is 100% at 0.494 FPs per slice and 8.75 FPs per patient and for the 20×20 cell templates, detection sensitivity is 86.66% with the level of 0.452 FPs per slice and 6.25 FPs per patient. © 2008 Springer Science+Business Media, LLC.
Computer network optimization using genetic algorithm
(2006) Akbulut, Olcay; Osman, Onur; Uçan, Osman N.
In this paper, Genetic Algorithm (GA) is proposed as optimization software to find the shortest path of various computer networks. It deals with different method concerning the placement routers, routes the packages. The genetic based algorithm defines an optimum way when a computer network system is constructed. Genetic Algorithm gives better results regarding other classical methods as the number of nodes of the network increases.
Concatenation of space-time block codes and turbo trellis coded modulation (ST-TTCM) over Rician fading channels with imperfect phase
(2004) Uçan, Osman N.; Osman, Onur
In this paper, the performance of space time-turbo trellis coded modulation (ST-TTCM) is evaluated over Rician and Rayleigh fading channels with imperfect phase. We modify Baum-Welch (BW) algorithm to estimate the fading and phase jitter parameters for multi-antenna configurations. Thus, we assume that the channel parameters change slower than carrier frequency. We know that, at high data rate transmissions over wireless fading channels, space-time block codes (STBC) provide the maximal possible diversity advantage. Here, the combined effects of the amplitude and the phase of the received signal are considered, each one modelled by Rician and Tikhonov distributions, respectively. We investigate space time-turbo trellis coded modulation (ST-TTCM) for 8-PSK for several Rician factor K and phase distortion factor ?. Thus, our results reflect the degradations both due to the effects of the fading on the amplitude and phase noise of the received signal while the channel parameters are estimated by BW algorithm. Copyright © 2004 John Wiley & Sons, Ltd.
Efficient estimation of osteoporosis using Artificial Neural Networks
(2007) Lemineur, Gerald; Harba, Rachid; Kılıç, Niyazi; Uçan, Osman N.; Osman, Onur; Benhamou, Laurent
In this communication, Artificial Neural Network (ANN) is applied to discriminate osteoporotic fracture and control cases in a group of 304 patients. ANN is one of the popular methods in optimization of complex engineering problems compared to the classical statistical methods. In our study group, we consider some parameters as inputs: three bone densitometry parameters (HMD) (Femoral neck BMD, Total Body BMD and L2L4 spine BMD), three fractal parameters [1,5] (Hmin. Hmean, Hmax), and age of the patient. We studied three ANN structures with various inputs and hidden neurons. We have reached up to 81.66% correct classification. In comparison we have tested a classical discriminant analysis (Mahalanobis-Fisher) and we only obtained 72% of correct classification. We can conclude that ANN is one of the promising methods in the diagnosis of osteoporosis. ©2007 IEEE.
Forward modeling with forced neural networks for gravity anomaly profile
(2007) Osman, Onur; Albora, A. Muhittin; Uçan, Osman N.
In this paper, we introduce a new method called Forced Neural Network (FNN) to find the parameters of the object in geophysical section respect to gravity anomaly assuming the prismatic model. The aim of the geological modeling is to find the shape and location of underground structure, which cause the anomalies, in 2D cross section. At the first stage, we use one neuron to model the system and apply back propagation algorithm to find out the density difference. At the second level, quantization is applied to the density differences and mean square error of the system is computed. This process goes on until the mean square error of the system is small enough. First, we use FNN to two synthetic data, and then the Sivas - Gürün basin map in Turkey is chosen as a real data application. Anomaly values of the cross section, which is taken from the gravity anomaly map of Sivas - Gürün basin, are very close to those obtained from the proposed method. © International Association for Mathematical Geology 2007.
Iterative cellular image processing algorithm
(2003) Osman, Onur; Uçan, Osman N.; Albora, A. Muhittin
In this paper, a new iterative image processing algorithm is introduced and denoted as “iterative cellular image processing algorithm” (ICIPA). The new unsupervised iterative algorithm uses the advantage of stochastic properties and neighborhood relations between the cells of the input image. In ICIPA scheme; first regarding to the stochastic properties of the data, all possible quantization levels are determined and then 2D input image is processed using a function, based on averaging and neighborhood relationship, and after that a parameter C is assigned to each cell. Then Gaussian probability values are mapped to each cell regarding to all possible quantization levels and the attended value C. A maximum selector defines the highest probability value for each cell. In the case of complex data, first iteration output is fed into input till a sufficient output is found. We have applied ICIPA algorithm to various synthetic examples and then a real data, the ruins of Hittite Empire. Satisfactory results are obtained. We have observed that de-noising property of our scheme is the best in the literature. It is interesting that the corrupted data with Additive White Gaussian Noise (AWGN) up to 97% ratio, can be de-noised by using our proposed ICIPA algorithm.
Lung nodule diagnosis using 3D template matching
(2007) Osman, Onur; Özekes, Serhat; Uçan, Osman N.
In this paper, to utilize the third dimension of Computed Tomography, regions of interest (ROI) slices were combined to form 3D ROI image and a 3D template was determined to find the structures with similar properties of nodules. Convolution of 3D ROI image with the proposed template strengthens the shapes similar to the template and weakens the other ones. False-positive (FP) per nodule and per slice versus diagnosis sensitivity were obtained. The Computer Aided Diagnosis system achieved 100% sensitivity with 0.83 FP per nodule and 0.46 FP per slice, when the nodule thickness was greater than or equal to 5.625 mm. © 2006 Elsevier Ltd. All rights reserved.
Mammographic mass classification using wavelet based support vector machine
(2009) Görgel, Pelin; Sertbaş, Ahmet; Kılıç, Niyazi; Uçan, Osman N.; Osman, Onur
In this paper, we investigate an approach for classification of mammographic masses as benign or malign. This study relies on a combination of Support Vector Machine (SVM) and wavelet-based subband image decomposition. Decision making was performed in two stages as feature extraction by computing the wavelet coefficients and classification using the classifier trained on the extracted features. SVM, a learning machine based on statistical learning theory, was trained through supervised learning to classify masses. The research involved 66 digitized mammographic images. The masses were segmented manually by radiologists, prior to introduction to the classification system. Preliminary test on mammogram showed over 84.8% classification accuracy by using the SVM with Radial Basis Function (RBF) kernel. Also confusion matrix, accuracy, sensitivity and specificity analysis with different kernel types were used to show the classification performance of SVM.
Performance of multilevel turbo codes with group partitioning over satellite channels
(2005) Osman, Onur; Uçan, Osman N.; Odabasıoğlu, Niyazi
A new type of turbo codes called multilevel turbo codes (ML-TC) which employ a new partitioning technique (group partitioning) in order to improve channel error performance. The underlying basis for multilevel coding is partitioning a signal set into several levels and encoding each level separately through the respective layer of the encoder. Initial partitioning levels are very important for the performance of ML-TC schemes. Group partitioning maximises the Euclidean distance of these levels, and it provides additional bit error performance augmentation of 0.6 dB for the range of signal-to-noise ratio (SNR) under study. The ML-TC system contains multiple turbo encoder/decoder blocks in its architecture. The parallel input data sequences are encoded by the multilevel scheme and mapped to any modulation type, such as MPSK, MQAM, etc. Then, for the purpose of performance analysis, these modulated signals are passed through narrowband fading channels. At the receiver side, the input sequence of the first level is estimated by the first turbo decoder block. Subsequently, the other input sequences of other levels are computed using the estimated input bit streams of the respective previous levels. Following that process, the performance of the proposed system is investigated in various fading channels. As an example, 4-PSK two-level turbo codes are simulated over AWGN, Rician and Rayleigh channels for 100 and 256 frame sizes. ML-TC simulation results display coding gains of up to 4.5 dB, when compared to those of 8-PSK turbo trellis-coded modulation. Therefore, it is concluded that satisfactory performance is achieved in ML-TC systems for all SNR values in various fading environments.
Performance of multilevel-turbo codes with blind/non-blind equalization over WSSUS multipath channels
(2006) Uçan, Osman N.; Buyukatak, Kenan; Göse, Ersin; Osman, Onur; Odabaşıoğlu, Niyazi
In this paper, in order to improve error performance, we introduce a new type of turbo codes, called 'multilevel-turbo codes (ML-TC)' and we evaluate their performance over wide-sense stationary uncorrelated scattering (WSSUS) multipath channels. The basic idea of ML-TC scheme is to partition a signal set into several levels and to encode each level separately by a proper component of the turbo encoder. In the considered structure, the parallel input data sequences are encoded by our multilevel scheme and mapped to any modulation type such as MPSK, MQAM, etc. Since WSSUS channels are very severe fading environments, it is needed to pass the received noisy signals through non-blind or blind equalizers before turbo decoders. In ML-TC schemes, noisy WSSUS corrupted signal sequence is first processed in equalizer block, then fed into the first level of turbo decoder and the first sequence is estimated from this first Turbo decoder. Subsequently, the other following input sequences of the frame are computed by using the estimated input bit streams of previous levels. Here, as a ML-TC example, 4PSK 2 level-turbo codes (2L-TC) is chosen and its error performance is evaluated in WSSUS channel modelled by COST 207 (Cooperation in the field of Science & Technology, Project #207). It is shown that 2L-TC signals with equalizer blocks exhibit considerable performance gains even at lower SNR values compared to 8PSK-turbo trellis coded modulation (TTCM). The simulation results of the proposed scheme have up to 5.5 dB coding gain compared to 8PSK-TTCM for all cases. It is interesting that after a constant SNR value, 2L-TC with blind equalizer has better error performance than non-blind filtered schemes. We conclude that our proposed scheme has promising results compared to classical schemes for all SNR values in WSSUS channels. Copyright © 2005 John Wiley & Sons, Ltd.
Performance of the systematic distance-4 codes over fading channels
(2008) Uçan, Osman N.; Acarer, Tayfun; Karaçuha, Ertuğrul; Osman, Onur; Altay, Gökmen; Yalçın, Şenay
A new binary systematic linear block code construction technique, called as Systematic Distance-4 (SD-4) codes that generates all the optimal size Hamming distance-4 codes, is recently proposed. In this paper, we evaluate the error performance of some of the SD-4 codes to measure the coding gain of the codes over Rician and Rayleigh fading channels. We also compared the results with the error performance of the same size extended Hamming codes, which is the subset of SD-4 codes.
Performance of transmit diversity-Turbo Trellis Coded Modulation (TD-TTCM) over genetically estimated WSSUS mimo channels
(Fachverlag Schiele und Sohn GmbH, 2004) Göse, Ersin; Pastacı, Halit; Uçan, Osman N.; Buyukatak, Kenan; Osman, Onur
This paper presents the performance of Transmit Diversity-Turbo Trellis Coded Modulation (TD-TTCM) over Wide Sense Stationary Uncorrelated Scattering (WSSUS), Multiple-lnput-Multiple-Output (MIMO) channels. To achieve high bandwidth efficiency and/or high power efficiency, in TD-TTCM, binary input sequence is passed through a turbo trellis encoder, mapped to 8-PSK and then fed into a new transmit diversity (TD) scheme for high data transmission over wireless fading channels. At the receiver side, the distorted multi-path signals are received by multiple receive antennas. WSSUS MIMO channel parameters and modulated signals are estimated by using Genetic Algorithm (GA) and an iterative combiner, respectively. Then they are taken as an input to Turbo trellis decoder. Here, TD-TTCM and its efficient implementations are discussed and simulation results are presented.
Performance of turbo decoding for time-varying multipath channels
(Institute of Electrical and Electronics Engineers Inc., 2003) Göse, Ersin; Büyükatak, Kenan; Osman, Onur; Uçan, Osman N.; Pastacı, Halit
In this paper, the performances of turbo coded signals are investigated over a new channel model, denoted as wide-sense stationary uncorrelated scattering (WSSUS) multipath channel. Digital transmission through WSSUS channel model with additive white Gaussian noise (AWGN) introduced at the receiver. For observing the performance of the turbo decoder, COST207 (Cooperation in the field of Science & Technology, Project 207) models (statistical channel models) based on WSSUS channel model are considered. Before decoding, the inverse filtering applied to equalize the corrupted data. LMS, RLS and Kalman algorithms are used for estimating inverse filter coefficients. Turbo decoder and its efficient implementation are discussed, and simulation results are presented. © 2003 IEEE.
A preliminary study on computerized lesion localization in MR mammography using 3D nMITR maps, multilayer cellular neural networks, and fuzzy c-partitioning
(John Wiley and Sons Ltd, 2008) Ertaş Gökhan; Gülçür, Halil Özcan; Tunacı, Mehtap; Osman, Onur; Uçan, Osman N.
Cellular neural networks (CNNs) are massively parallel cellular structures with learning abilities. They can be used to realize complex image processing applications efficiently and in almost real time. In this preliminary study, we propose a novel, robust, and fully automated system based on CNNs to facilitate lesion localization in contrast-enhanced MR mammography, a difficult task requiring the processing of a large number of images with attention paid to minute details. The data set consists of 1170 slices containing one precontrast and five postcontrast bilateral axial MR mammograms from 39 patients with 37 malignant and 39 benign mass lesions acquired using a 1.5 Tesla MR scanner with the following parameters: 3D FLASH sequence, TR/TE 9.80/4.76 ms, flip angle 25°, slice thickness 2.5 mm, and 0.625×0.625 mm2 in-plane resolution. Six hundred slices with 21 benign and 25 malignant lesions of this set are used for training the CNNs; the remaining data are used for test purposes. The breast region of interest is first segmented from precontrast images using four 2D CNNs connected in cascade, specially designed to minimize false detections due to muscles, heart, lungs, and thoracic cavity. To identify deceptively enhancing regions, a 3D nMITR map of the segmented breast is computed and converted into binary form. During this process tissues that have low degrees of enhancements are discarded. To boost lesions, this binary image is processed by a 3D CNN with a control template consisting of three layers of 11×11 cells and a fuzzy c -partitioning output function. A set of decision rules extracted empirically from the training data set based on volume and 3D eccentricity features is used to make final decisions and localize lesions. The segmentation algorithm performs well with high average precision, high true positive volume fraction, and low false positive volume fraction with an overall performance of 0.93±0.05, 0.96±0.04, and 0.03±0.05, respectively (training: 0.93±0.04, 0.94±0.04, and 0.02±0.03; test: 0.93±0.05, 0.97±0.03, and 0.05±0.06). The lesion detection performance of the system is quite satisfactory; for the training data set the maximum detection sensitivity is 100% with false-positive detections of 0.28/lesion, 0.09/slice, and 0.65/case; for the test data set the maximum detection sensitivity is 97% with false-positive detections of 0.43/lesion, 0.11/slice, and 0.68/case. On the average, for a detection sensitivity of 99%, the overall performance of the system is 0.34/lesion, 0.10/slice, and 0.67/case. The system introduced does not require prior information concerning breast anatomy; it is robust and exceptionally effective for detecting breast lesions. The use of CNNs, fuzzy c -partitioning, volume, and 3D eccentricity criteria reduces false-positive detections due to artifacts caused by highly enhanced blood vessels, nipples, and normal parenchyma and artifacts from vascularized tissues in the chest wall due to oversegmentation. We hope that this system will facilitate breast examinations, improve the localization of lesions, and reduce unnecessary mastectomies, especially due to missed multicentric lesions and that almost real-time processing speeds achievable by direct hardware implementations will open up new clinical applications, such as making feasible quasi-automated MR-guided biopsies and acquisition of additional postcontrast lesion images to improve morphological characterizations. © 2008 American Association of Physicists in Medicine.

Yazar "Uçan, Osman N." seçeneğine göre listele

Sayfa Başına Sonuç

Sıralama seçenekleri