Search

25th Signal Processing and Communications Applications Conference (SIU) -- MAY 15-18, 2017 -- Antalya, TURKEY WOS: 000413813100213 Superpixel based methods have recently shown success in scene segmentation and labeling. In scene labeling, a superpixel algorithm is used first to segment the image into visually consistent small regions; then several feature descriptors are computed and classification is performed for each superpixel. In this paper, Kernel Codebook Encoding (KCB) of superpixel features is proposed. In KCB feature vectors are mapped to multiple codewords in a soft manner, instead of the usual hard quantization. The weights assigned to the codewords are determined by a kernel distance function. KCB method is used for encoding of SIFT features in SuperParsing image parsing algorithm. The developed approach is tested on the SIFT Flow dataset consisting of 2,688 images and 33 classes, and achieves 2.7% increase in parsing accuracy over SuperParsing. Turk Telekom, Arcelik A S, Aselsan, ARGENIT, HAVELSAN, NETAS, Adresgezgini, IEEE Turkey Sect, AVCR Informat Technologies, Cisco, i2i Syst, Integrated Syst & Syst Design, ENOVAS, FiGES Engn, MS Spektral, Istanbul Teknik Univ

Traveling Salesman Problem (TSP) is the problem of finding a minimum distance tour of cities beginning and ending at the same city and that each city are visited only once. As the number of cities increases, it is difficult to find an optimal solution by exact methods in a reasonable duration. Therefore, in recent five decades many heuristic solution methods inspired of nature and biology have been developed. In this paper, a new metaheuristic method inspired of the by-passing the obstacle strategy of blind mole rats living in their individual tunnel systems under the soil is designed for solving TSP. The method is called as Blind Mole-rat Algorithm. The proposed algorithm is tested on different size of symmetric TSP problems and the results are compared to the best known results. Initial test results are promising although proposed metaheuristic is not yet competitive enough among other algorithms in the literature.Keywords: Traveling salesman problem, Combinatorial optimization, Metaheuristic, Blind mole-Rat algorithm, Gezgin Satıcı Problemi (GSP), başlangıç ve bitiş şehirleri aynı olan ve her şehrin sadece bir kez ziyaret edildiği minimum mesafeli turu bulma problemidir. Şehir sayısı arttıkça, kesin yöntemler ile kabul edilebilir sürelerde bir optimum çözüm bulunması zordur. Bu nedenle, son elli yılda GSP’nin çözümü için doğadan ve biyolojiden esinlenen birçok meta-sezgisel yöntem geliştirilmiştir. Bu çalışmada, toprak altındaki bireysel tünel sistemlerinde yaşayan kör farelerin toprak altındaki engelleri geçme stratejisinden esinlenilerek GSP’nin çözümü için yeni bir meta-sezgisel tasarlanmıştır. Geliştirilen yönteme Kör Fare Algoritması adı verilmiştir. Bu yeni sezgisel ile farklı boyutlardaki simetrik test veri setleri için deneyler yapılmış ve sonuçları bilinen en iyi sonuçlar ile kıyaslanmıştır. Önerilen meta-sezgisel henüz literatürdeki diğer algoritmalarla yarışabilecek düzeyde olmamasına rağmen, başlangıç test çözümlerinin umut verici olduğu söylenebilir.Anahtar Kelimeler: Gezgin satıcı problemi, Kombinatoryel eniyileme, Meta-Sezgisel, Kör fare algoritması

With rapid advances of powerful image processing softwares, it has become easier to commit a forgery by damaging originality of an image. Copy move forgery is one of the most widely used methods of image forgery. In this paper, a copy-move image forgery detection technique is proposed which use together LBP (Local Binary Pattern) and DCT (Discrete Cosine Transform). Firstly, we divide the image into overlapping blocks and then we convert each image block to LBP domain. We have derived the feature matrix of the image by applying DCT to the image blocks which is converted to LBP domain. We lexicographically sorted the feature matrix and after this step, similar image blocks are sequentially sorted. It is aimed that minimising the false positives and deriving of local features which is less sensitive to varying lighting conditions by applying DCT to the image blocks which is converted to LBP domain. Experimental results show that when used together LBP and DCT demonstrate the better results with fewer false positives according to individual usage in copy-move image forgery detection.

Compressed sensing theory shows that any signal which is defined as sparse in a given domain can be reconstructed using fewer linear projections instead of using all Nyquist-rate samples. In this paper, we investigate basis pursuit, matching pursuit, orthogonal matching pursuit and compressive sampling matching pursuit algorithms, which are basic compressed sensing based algorithms, and present performance curves in terms of mean squared error for various parameters including signal-to-noise ratio, sparsity and number of measurements with regard to mean squared error. In addition, accuracy of estimation performances has been supported with theoretical lower bounds (Cramer-Rao lower bound and deterministic lower mean squared error). Considering estimation performances, compressive sampling matching pursuit yields the best results unless the signal has a non-sparse structure.

In this paper, we proposed a speech signal denoising method using common vector approach (CVA) for very noisy signals. Based on CVA, classes are constructed from windowed speech samples/vectors according to their characteristics and denoising is performed on difference vectors. Speech signal is then reconstructed using denoised patches. Size of windows and amount of overlapping are important parameters in this method, affecting the performance/noise reduction ratio. CVA results are compared against the results of Geometric Approach and Magnitude Squared Spectrum methods according to well-known performance measures. Experimental results show that the proposed common vector based method is superior to other techniques for very noisy speech signals. Method is open to enhancement with different parameters and noise estimation methods.

This paper presents a new partial optimization approach for the least significant bit (LSB) data hiding technique that can be used for protecting any secret information or data. A deterioration effect of as little as possible in an image is intended using the LSB data hiding technique and this is well realized utilizing the proposed partial optimization approach achieving the same data embedding bit rates. In the proposed approach, all of the image pixels are classified into 8 regions and then the 8 distinct ordering codings are applied to each region by the developed partial optimization encoder. Thus, the most effective outcome that has been obtained from the 8 regions means that the number of the altered bits is kept minimized. Hence, the minimal values that have been attained from the 8 regions enable decoding that ensures relatively small distortions on the extracted cover image.

22nd IEEE Signal Processing and Communications Applications Conference (SIU) -- APR 23-25, 2014 -- Karadeniz Teknik Univ, Trabzon, TURKEY WOS: 000356351400252 In this paper, an prediction speed method of hand open-close by using the Artificial Neural Network (ANN) surface electromyography (sEMG) signal is presented. The first step of this method is to analyze sEMG signal detected from the subject's right upper forearm and extract features using the mean absolute value (MAV), the root mean square (RMS), the variance (VAR), the standart deviation (STD), the median frekans of power spectrum (MDF), the mean frekans of PS (MNF), the maximum frekans of PS (MAXF). The second step is to import the feature values into an ANN to identify the speed of hand open-close (SHOC). Based on the results of experiments, it is concluded that this method is effective in prediction of SHOC. IEEE, Karadeniz Tech Univ, Dept Comp Engn & Elect & Elect Engn

The Microsoft Kinect enables real time 3D reconstruction of an indoor environment with low computational power by producing depth measurements with high resolution. Yet, the measured depth images are not in perfect quality because of the reflectance properties of the objects inside the scene. In this paper, we propose a novel depth prediction algorithm that is based on fitting a polynomial on the unknown depth locations. The parameters of the polynomials are found using the features and the depth measures of the pixels with known depth value. We conducted several experiments on different images and achieved predicting the depth of unknown locations with low error.

21st Signal Processing and Communications Applications Conference (SIU) -- APR 24-26, 2013 -- CYPRUS WOS: 000325005300016 In literature video coding efficiency of H. 264 standard has been improved using decoder-side super-resolution (SR). In this paper SR estimation is used during encoding in place of skip/direct mode prediction of H. 264 standard and coding performance is optimized. First P-and B-frames are downsampled and H. 264 encoded at lower resolution (LR). Then reconstructed LR frames are super-resolved using decoder-side motion estimation. Also bicubic interpolation of the LR frame is used as an alternative to SR reconstruction. A rate-distortion optimal mode selection algorithm determines for each macroblock which of the two reconstructions to use as skip/direct mode prediction. Simulations indicate an average of 1.36 dB PSNR improvement or 30.3% bitrate reduction at low bitrates when compared to H. 264 standard. Average PSNR gains reach as high as 3.63 dB depending on the video content.

Digital images are distorted by impulsive noise (IN) which caused by image sensors and communication channels in many practical applications. This noise may cause miscalculation of sharpness values in image fusion applications. In this paper, traditional criterion functions and frequency selective median filter (FSWM) based criterion function are evaluated for the fusion of multi-focus images in the presence of IN. Experimental results show that the FSWM filter is better than the other focus measures.

2009 International Conference on Application of Information and Communication Technologies, AICT 2009 -- 14 October 2009 through 16 October 2009 -- Baku -- 79447 In the framework of the electrocardiography (ECG) signals, this paper describes an original approach to identify heart beat morphologies and to detect R-wave events. Examining the localization of R wave carefully may lead to an accurate diagnosing for the heart abnormalities. Several techniques have been conducted in the area of arrhythmia detection using classical methods of classification In this work, the proposed approach is based on a "polynomial curve fitting (PCF)" rule evaluated using a decision function in a local moving-window procedure (inspired by [10]). The data set used were taken from MIT-BIH Arrhythmia database, as well as several other international ECG signal databases. Most of the localization of R wave detected using this novel method with full degree of confidence. And heart rate was evaluated using this novel method with a high degree of confidence. ©2009 IEEE.

This paper presents a novel method for 3D facial feature extraction. In this method, 3D Discrete Cosine Transform (DCT) is used to extract features. Before the feature extraction, faces are aligned with respect to nose tip and then registered two times: according to average nose and average face. Then the coefficients of 3D transformation are calculated. The most discriminant 3D transform coefficients are selected as feature vector where the ratio of within-class variance and between-class variance is used as selection measure. To further improve the class discrimination, Linear Discriminant Analysis (LDA) is applied to feature vectors before the matching procedure. We also investigate the proposed approach by using 3D Discrete Fourier Transform instead of 3D DCT for comparison. Selected and discriminated features are matched by using the Nearest Neighbor classifier. Recognition experiments are realized on 3D RMA face database. The proposed method shows the recognition rate above 99% for 3D DCT based features.