Yazar "Polat, Ediz" seçeneğine göre listele
Listeleniyor 1 - 12 / 12
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Automatic Detection and Classification of Non Metallic Inclusions in Steel via Image Processing(Ieee, 2014) Aydilek, Huseyin; Polat, EdizNon-metallic chemical components (silicate, sulfide, alumina, globular oxide) originated from production process cause inclusions in steel. These inclusions are one of the important factors that directly affect quality of steel. Before raw steel is used, the detection and classification of the inclusions according to international standards ASTM E45 will prevent faults that may occur during or after production. In this paper, an accurate, fast and reliable image processing based system that automatically detects and classifies non metallic inclusions in steel according to ASTM E45 standards is developed. The system is tested for detection and classification of Type A inclusions and successful results are obtained.Öğe A color image segmentation approach for content-based image retrieval(Pergamon-Elsevier Science Ltd, 2007) Özden, Mustafa; Polat, EdizThis paper describes a new color image segmentation method based on low-level features including color, texture and spatial information. The mean-shift algorithm with color and spatial information in color image segmentation is in general successful, however, in some cases, the color and spatial information are not sufficient for superior segmentation. The proposed method addresses this problem and employs texture descriptors as an additional feature. The method uses wavelet frames that provide translation invariant texture analysis. The method integrates additional texture feature to the color and spatial space of standard mean-shift segmentation algorithm. The new algorithm with high dimensional extended feature space provides better results than standard mean-shift segmentation algorithm as shown in experimental results. (c) 2006 Pattern Recognition Society. Published by Elsevier Ltd. All rights reserved.Öğe Color image segmentation using mean shift filtering and wavelet frames(2005) Özden, Mustafa; Polat, Ediz; Tuna, EyüpThis paper describes a new color image segmentation method based on low-level features including color and texture. The mean-shift algorithm with color and spatial information in color image segmentation is in general successful, however, in some cases, the color and spatial information are not sufficient for superior segmentation. Recently, the discrete wavelet transform (DWT) has become a popular approach for texture feature extraction. The proposed method uses wavelet frames that provide translation invariant texture analysis. The method also integrates an additional texture feature to the traditional mean shift segmentation algorithm with color and spatial space. Experimental results show that the algorithm gives satisfactory results. © 2005 IEEE.Öğe The Effect of Dictionary Learning Algorithms on Super-resolution Hyperspectral Reconstruction(Ieee, 2015) Simsek, Murat; Polat, EdizThe spatial resolutions of hyperspectral images are generally lower due to imaging hardware limitations. Super-resolution algorithms can be applied to obtain higher resolutions. Many algorithms exist to achieve super-resolution hyperspectral images from low resolution images acquired in different wavelengths. One of the popular algorithms is sparse representation-based algorithms that employ dictionary learning methods. In this study, a comparative framework is developed to investigate which dictionary learning algorithm leads to better super-resolution images. In order to achieve that, K-SVD and ODL dictionary learning algorithms are employed for comparison. A sparse representation-based algorithm G-SOMP+ is used for hyperspectral super-resolution reconstruction. The experimental results show that ODL algorithm outperforms K-SVD in terms of both reconstruction quality and processing times.Öğe Feature Selection for Multi-SVM Classifiers in Facial Expression Classification(Ieee, 2008) Gunes, Turan; Polat, EdizFacial expressions are non-verbal signs that play an important role between interpersonal communications and also they are the most effective way to describe human emotions. The correct and fast extraction/recognition of the facial expressions is an ongoing research area for computer vision. In this study, the effects of feature selection on the classification of seven different facial expressions (anger, disgust, fear, joy, neutral, sadness, and surprise) are analyzed. To this end, the features for each expression were extracted using Gabor filters from the facial images and selected the best ones using two different feature categories. In the first category, the features that they do exist in one class but do not exist in all other classes have been selected. In the second category, the features that they represent the one class have been selected. These selected features were employed for expression classification using Support Vector Machines (SVMs). Comparative results are given in terms of types for selecting features, feature selection algorithms and the approaches used for the multi-classification. The best results were obtained using RFE algorithm with first type of feature selection and using one-vs-rest approach for training. It was also demonstrated that the feature selection has a positive effect to increase the classification performance comparing when the results were observed with no feature selection.Öğe Feature selection in facial expression analysis and its effect on multi-SVM classifiers(Gazi Univ, Fac Engineering Architecture, 2009) Gunes, Turan; Polat, EdizFacial expressions are non-verbal signs that play important role to provide complete meaning in human communication. While humans can easily comprehend the facial expressions, it is not valid for the computers, thus the researchers are still working on developing reliable facial expression recognition systems. In this research, the analysis of 7 different human facial expressions (anger, disgust, fear, happiness, neutral, sadness and surprise) is performed from human facial images. For this purpose, the features for every facial expression are extracted using Gabor filters. The feature sets that best represent the facial expressions are obtained using different feature selection algorithms. The effects of selected feature sets on the multi-class Support Vector Machine (SVM) classifiers are investigated and a comparative evaluation for classification results is given for each algorithm. For the multi-class classification, the SVM classifier is used with three different approaches including One-Vs-One, One-Vs-Rest and Multi-class SVM. It is also shown that classification rates are increased when the selected features are used.Öğe Image segmentation using color and texture features(2005) Özden, Mustafa; Polat, EdizThis paper describes a new color image segmentation method based on low-level features including color, texture and spatial information. The mean-shift algorithm with color and spatial information in color image segmentation is in general successful, however, in some cases, the color and spatial information are not sufficient for superior segmentation. The proposed method addresses this problem and employs texture as an additional feature. The method uses wavelet frames that provide translation invariant texture analysis. The method integrates additional texture feature to the color and spatial space of standard mean shift segmentation algorithm. The new algorithm with high dimensional extended feature space provides better results than standard mean shift segmentation algorithm as shown in experimental results.Öğe A nonparametric adaptive tracking algorithm based on multiple feature distributions(Ieee-Inst Electrical Electronics Engineers Inc, 2006) Polat, Ediz; Özden, MustafaThis paper presents an object tracking framework based on the mean-shift algorithm, which is a nonparametric technique that uses statistical color distribution of objects. Tracking objects through highly similar-colored background is one of the problems that need to be addressed. In various cases where object and background color distributions are very similar, the color distribution obtained from single frame alone is not sufficient to track objects reliably. To deal with this problem, the proposed algorithm utilizes an adaptive statistical background and foreground modeling to detect the change due to motion using kernel density estimation techniques based on multiple recent frames. The use of multiple frames supplies more information than single frame and thus it provides more accurate modeling of both background and foreground. In addition to color distribution, this statistical multiple frame-based motion representation is integrated into a modified mean-shift algorithm to create more robust object tracking framework. The use of motion distribution provides additional discriminative power to the framework. The superior performance with quantitative results of the framework has been validated using experiments on synthetic and real sequence of images.Öğe Sparse Representation-based Dictionary Learning Methods for Hyperspectral Super-Resolution(Ieee, 2016) Simsek, Murat; Polat, EdizDue to hardware limitations, hyperspectral imagery has low spatial resolution. It can be obtained super-resolution hyperspectral imagery by means of sparse representation-based methods that are designed for improving spatial resolution. In this paper, the effect of sparse representation-based dictionary learning algorithms including K-SVD, ODL and Bayes on obtaining superresolution images with low error and high quality has been investigated. The method with best results has been identified.Öğe Stationary Aircraft Detection From Satellite Images(Istanbul Univ, Fac Engineering, 2012) Polat, Ediz; Yıldız, CihatSatellite image analysis is an important research area in the field of image processing. Detection and recognition of regions and objects from satellite images find many useful civil applications such as detection of buildings, roads, bridges and other man-made objects as well as land plant classification. On the other hand, the detection of stationary aircrafts in airports can be strategically important in military applications. In this study, a learning-based system that detects stationary aircrafts in satellite images obtained from Google Earth is developed. The features that emphasize the geometric structure of an aircraft are determined using 2D Gabor filter. The aircraft detection is performed using Support Vector Machines (SVM) classification method. The SVM is a supervised learning method that analyzes data and recognizes patterns for classification The SVM takes a set of input data (a vector consists of Gabor filter output of images) and predicts the one of two classes (aircraft or non-aircraft). The performance of the system is demonstrated using satellite images collected from airports in Europe and United States.Öğe A video-based eye pupil detection system for diagnosing bipolar disorder(Tubitak Scientific & Technical Research Council Turkey, 2013) Akinci, Gokay; Polat, Ediz; Kocak, Orhan MuratEye pupil detection systems have become increasingly popular in image processing and computer vision applications in medical systems. In this study, a video-based eye pupil detection system is developed for diagnosing bipolar disorder. Bipolar disorder is a condition in which people experience changes in cognitive processes and abilities, including reduced attentional and executive capabilities and impaired memory. In order to detect these abnormal behaviors, a number of neuropsychological tests are also designed to measure attentional and executive abilities. The system acquires the position and radius information of eye pupils in video sequences using an active contour snake model with an ellipse-fitting algorithm. The system also determines the time duration of the eye pupils looking at certain regions and the duration of making decisions during the neuropsychological tests. The tests are applied to 2 different groups consisting of people with bipolar disorder (bipolar group) and people without bipolar disorder (control group) in order to mathematically model the people with bipolar disorder. The mathematical modeling is performed by using the support vector machines method. It is a supervised learning method that analyzes data and recognizes patterns for classification. The developed system acquires data from the being tested and it classifies the person as bipolar or nonbipolar based on the learned mathematical model.Öğe Yüz ifade analizinde öznitelik seçimi ve çoklu SVM sınıflandırıcılarına etkisi(2009) Güneş, Turan; Polat, EdizYüz ifadeleri, insan ilişkilerinde anlam bütünlüğünün sağlanması için büyük rol oynayan, sözlü olmayan işaretlerdir. İnsanoğlu yüz ifadelerini kavramada herhangi bir zorluk çekmezken, bu durum makineler için geçerli olmayıp, halen güvenilir ifade tanıma sistemleri üzerinde araştırmalar yapılmaktadır. Bu çalışmada, insanın içinde bulunabileceği 7 ifade durumunun (öfke, iğrenme, korku, mutluluk, ifadesizlik, üzüntü ve şaşkınlık) analizi gerçekleştirilmiştir. Bu amaçla, her bir ifade için alınan sabit görüntülerin öznitelikleri Gabor filtreleri kullanılarak çıkartılmış ve farklı öznitelik seçme algoritmaları kullanılarak ifadeleri temsil eden en iyi öznitelik kümeleri oluşturulmuştur. Seçilen öznitelik kümelerinin çoklu SVM (Support Vector Machines-Destek Vektör Makineleri) sınıflandırıcıları üzerindeki etkileri incelenmiş ve sınıflandırma doğruluklarının kullanılan öznitelik seçme algoritmalarına göre nasıl değiştiği karşılaştırmalı olarak değerlendirilmiştir. Çoklu sınıflandırma yapılması amacıyla SVM, One-Vs-One, One-Vs-Rest ve MC-SVM olmak üzere 3 farklı yaklaşım ile birlikte kullanılmıştır. Ayrıca öznitelik seçimi yapılmadan alınan sınıflandırma başarım sonuçları da incelendiğinde, öznitelik seçiminin sınıflandırma doğruluğunun artması yönünde genel olarak büyük etkisinin olduğu görülmüştür.
