Search

A means of transportation is the way in which an object, person, or service is transported from one place to another. Rail transportation occupies an important place in terms of cost and reliability. Most train accidents are caused by faults in railroad tracks. Detecting faults in railroad tracks is a difficult and time-consuming process compared to conventional methods. In this study, an artificial intelligence based model is proposed that can detect faults in railroad tracks. The dataset used in the study consists of defective and non-defective railroad images. The proposed model consists of foldable neural networks developed using the Tensorflow library. Softmax method was used as a classifier. An overall accuracy of 92.21% was achieved in the experiment.

Günümüzde, trafik hayatında seyreden araç yoğunluğu ciddi boyutlara ulaşmıştır. Bu nedenle, mevcut ulaşım ağlarının kullanım kapasitesi maksimum seviyelere çıkmakta ve trafik sıkışıklığına yol açmaktadır. Akıllı Ulaşım Sistemlerinin bir çözümü olan Görsel Trafik Gözetleme Sistemleri trafik sıkışıklığını azaltmak için kullanılan alternatif yöntemlerden biridir. Görsel Trafik Gözetleme Sisteminin temel görevlerinden biri; video veya görüntülerden algılanan araç türlerini doğru bir şekilde sınıflandırmaktır. Bu çalışma, Görsel Trafik Gözetleme Sisteminin araç türlerini sınıflandırma doğruluğunu arttıracak yeni yöntemler sunmayı amaçlamaktadır. Çoğu görüntü sınıflandırma doğruluğunu arttıran çalışmalarda geleneksel yöntemler kullanılırken bu çalışmada günümüzde trend olan mobil evrişimli sinir ağları (MCNN) iki farklı yaklaşımla ele alınmaktadır. İlk olarak, MobileNetv1 ve MobileNetv2 modelleri optimize edilerek İMobileNetv1 ve İMobileNetv2 yaklaşımları önerildi. İkinci olarak, bu önerilen MCNN yaklaşımları sadece özellik çıkarıcı olarak kullanıldığı ve elde edilen özelliklerin birleştirilmesi, seçilmesi ve sınıflandırılması gibi yöntemlerin birlikte kullanıldığı bir yaklaşım önerildi. Önerilen yaklaşımlarla yapılan sınıflandırma sonucunda, %85,05 oranında çok yüksek bir sınıflandırma başarısı elde edilmiştir.

Evrişimsel Sinir ağları (ESA), son yıllarda birçok çalışma tarafından özellik çıkarıcı olarak sıklıkla kullanılmaktadır. El ile çıkarılan farklı özellik çıkarma algoritmalarının aksine, etkileşim gerektirmeden otomatik olarak özellik çıkaran ESA’ların yardımıyla, birçok problem ve çalışma alanındaki başarımlar daha ileriye taşınmıştır. Bu çalışmada, farklı mimari özelliklere sahip olan ESA’ların yüz ifade analizi üzerindeki başarımları incelenmiştir. Öncelikle farklı ESA mimarileri tanıtılmış ve bu mimarilerin birbirlerinden farklılaştıkları kısımlar açıklanmıştır. FER2013 veri seti kullanılarak bütün ağ mimarileri üzerinde gerçekleştirilen eğitim ve doğrulama işlemleri sonucunda her bir mimariye ait başarım ve kayıp grafikleri sunulmuştur. Son olarak farklı mimarilerin yüz ifade analizi üzerindeki başarımlarının sebepleri tartışılmış ve gelecek çalışmalar için önerilerde bulunulmuştur.

Görüntü işleme yöntem ve tekniklerinin gün geçtikçe daha iyi sonuç vermesi, ekolojik dengenin duyarlılığı açısından önem arz etmektedir. Bu makale ekolojik dengenin temel unsuru olan çiçek görüntülerinin sınıflandırılmasını ele almaktadır. Son zamanlarda çiçek görüntüleri üzerinde derin öğrenme yöntemlerinin kullanımı artmıştır. Bu çalışmada, çiçek görüntülerinin sınıflandırılması için internette erişime açık olan veri seti kullanılmıştır. Veri seti 4326 görüntüden oluşmaktadır. Elde edilen veri kümesinde özellik çıkarımı için derin öğrenme modellerinden evrişimsel sini ağı (ESA) kullanılmıştır. ESA mimarilerinden AlexNet, VGG-16 ve VGG-19 modelleri kullanılmıştır. Üç modelinde ortak özelliği 1000 özellik veren tam bağlantılı katmana sahip olmalarıdır. Çiçek görüntülerinden elde edilen özellikler destek vektör makineleri (DVM) ile sınıflandırılmış ve elde edilen sonuçlar karşılaştırılmıştır. Karşılaştırma sonucunda en iyi sınıflandırma performansını VGG-16 mimarisi ile sağlanmıştır. Elde edilen sınıflandırma doğruluk oranı %86,56’dır. Sonraki aşamada ESA mimarilerinin son tam bağlantılı katmanından elde edilen 1000 özellik birleştirilerek 3000 özellik seti oluşturuldu. Ardından, özellik seçim yöntemlerinden; Maksimum Bilgi Katsayısı (MBK), Ridge regresyonu ve Özyinelemeli Özellik Eleme (ÖÖE) yöntemleri kullanılarak özellik sayısı 300’e düşürülmüştür. Özellik seçim yöntemleri ile çıkartılan en verimli özellikler DVM yöntemi ile yeniden sınıflandırılmıştır. Sınıflandırma başarı oranı yaklaşık %4,54 artarak %91,10 olmuştur. Bu çalışma ile çiçek görüntülerinin sınıflandırılmasında ESA mimarileri ile birlikte özellik seçim yöntemlerinin kullanımının etkili olduğu gözlemlenmiştir.

It is an important work area to determine realtime characteristics of roads where vehicles are in motion in critical areas where artificial intelligence is effectively used, such as driverless vehicles. The purpose of this article work is to present a deeper learning method that will allow a vehicle in motion to detect the direction of flow in the path. Convolutional Neural Networks (KSA) have been used as deep learning models for the determination of the direction of flow (YAY) in the study. The YAY-KSA model developed for flow direction detection is applied on 587 real road images in the CMU VASC image database. To compare the performances of the prepared model, Cifar model which is a common KSA model was applied on the same data. According to the classification results obtained, it was seen that the designed YAY-KSA model correctly determined flow direction at 80.1% level.

Kentleşmenin şehirlerin hızlı gelişmesine neden olması gelişen şehirlerin alt yapılarının takibi ve güncellenme gerekliliğinin analizini zorunlu kılmıştır. Özellikle ulaşım karmaşasının önüne geçmek için çeşitli trafik planlamaları yapmanın yanı sıra ulaşımın sağlandığı kara yolunun niteliğinin de yeterli seviyede olması gereklidir. Yolun yapısal kusurların ve çatlaklarının manuel görsel muayenesi verinin hacmi ve yapının boyutu nedeniyle çok zaman alan zahmetli bir süreçtir. Yoldaki çatlakların ve kusurların manuel olarak incelenmesi, yorgunluk, sorumsuz denetim, zayıf göz görme gibi bir dizi nedenlerden dolayı insan hatası nedeniyle yeterli seviyede değerlendirilememektedir. Yol kusurlarının belirlenmesi, sürücüler için önemli olmakla beraber yaya gibi tüm yol kullanıcılarının güvenliği ve konforunu sağlamak için çukurlar, hız tümsekleri vb. yol yüzeyi anormalliklerinin izlenmesi büyük önem taşımaktadır. Bu çalışmada yol yüzey kalitesinin izlenmesine ve sürücülere daha güvenli bir yol sunmak adına kullanışlı bir otomatik algılama sisteminin geliştirilmesine odaklanmaktadır. Veri seti kamera sayesinde alınan verilerle oluşturulmuştur. Verilerin önişleme fazı tamamlandıktan sonra VGG kullanılarak sonuçlar alınmış, kazanç ve kayıp grafikleri çizdirilmiş ve tahminler yapılmıştır.

This experimental study aimed to investigate a relationship between tinnitus and thicknesses of internal auditory canal and nerves in it. It was performed on brain magnetic resonance images of patients who consulted the ear, nose, and throat clinic with tinnitus complaint. Statistical hypothesis tests and classification experiments were performed on these data to find out structural differences in internal auditory channel components in patients with tinnitus after obtaining cross-sectional areas of nerves as thicknesses. Both the hypothesis tests and classification results showed that the thicknesses of nerves in tinnitus cases were different from those in normal cases. In particular, the hypothesis tests for the superior vestibular nerve and internal auditory channel showed the highest significance, indicating the relationship with tinnitus. The classification results indicated the possibility of classification for tinnitus identification, establishing a computer-assisted diagnostic system to help physicians.

Classifying detected software vulnerabilities is an important process. However, the metric values of security vectors are manually determined by humans, which takes time and may introduce errors stemming from human nature. These metrics are important because of their role in the calculation of vulnerability severity. It is necessary to use machine learning algorithms and data mining techniques to improve the quality and speed of vulnerability analysis and discovery processes. However, studies in this area are still limited. In this study, vulnerability vectors were estimated using the natural language processing techniques bag of words, term frequency–inverse document frequency, and n-gram for feature extraction together with various multiclass classification algorithms, namely Naive Bayes, decision tree, k-nearest neighbors, multilayer perceptron, and random forest. Our experiments using a large public dataset facilitate assessment and provide a standard-compliant prediction model for classifying software vulnerability vectors. The results show that the joint use of different techniques and classification algorithms is a promising solution to a multi-probability and difficult-to-predict problem. In addition, our study fills an important gap in its field in terms of the size of the dataset used and because it covers a vulnerability scoring system version that has not yet been extensively studied.

Coronavirus causes a wide variety of respiratory infections and it is an RNA-type virus that can infect both humans and animal species. It often causes pneumonia in humans. Artificial intelligence models have been helpful for successful analyses in the biomedical field. In this study, Coronavirus was detected using a deep learning model, which is a sub-branch of artificial intelligence. Our dataset consists of three classes namely: coronavirus, pneumonia, and normal X-ray imagery. In this study, the data classes were restructured using the Fuzzy Color technique as a preprocessing step and the images that were structured with the original images were stacked. In the next step, the stacked dataset was trained with deep learning models (MobileNetV2, SqueezeNet) and the feature sets obtained by the models were processed using the Social Mimic optimization method. Thereafter, efficient features were combined and classified using Support Vector Machines (SVM). The overall classification rate obtained with the proposed approach was 99.27%. With the proposed approach in this study, it is evident that the model can efficiently contribute to the detection of COVID-19 disease., Graphical abstract Image 1, Highlights • Chest data obtained from patients infected with the new Coronavirus (COVID-19) were used. • It was detected with deep learning models using COVID-19, normal, and pneumonia chest data. • The original dataset was restructured with the Fuzzy Color technique and two datasets were stacked. • Efficient features were selected by applying Social Mimic optimization to feature sets extracted from CNN models. • The efficient features obtained were combined, and classified with a success rate of 99.27% with SVM method.

Communication is the process of people transferring emotions, thoughts or information to the other party in various ways. One of the most effective ways of communication is language. Language is a communication tool that makes people's daily life easier and there are many hearing impaired people in our lives who cannot use this communication tool. Sign languages have been developed to facilitate the communication of hearing impaired people in society. There are specific sign languages varying according to the language of the countries. This study focuses on the Turkish sign language digits that are publicly available. Sign language is not known by all people of society. This situation causes communication disruptions in the social environments where hearing impaired people are present. A person who has not hearing impaired but cannot use sign language has the same problem. The aim of this study is to determine what people using sign language want to tell by using a deep learning architecture. For this purpose, the identification of digits in Turkish sign language has been realized by using the recently popular siamese neural network in this study. Siamese neural networks are a type of deep learning model that matches the same images in an image dataset. Using these networks, we have identified the digits used in Turkish sign language. The success rate of the matching was 98.16%. Consequently, siamese neural networks were found to be successful in identifying Turkish sign language digits with this study., İletişim, insanların duygu,düşünce veya bilgiyi çeşitli yollar kullanarak karşı tarafa aktarma sürecidir.İletişimde en etkili yollardan birisi ise dildir. Dil, insanların günlükhayatını kolaylaştıran bir iletişim aracıdır ve bu iletişim aracınıkullanamayan işitme engelli birçok insan vardır. İşitme engelli insanların,toplum içerisinde iletişimini kolaylaştırmak için işaret dilleri geliştirilmiştir.Her ülkenin kendi konuşma diline özgü işaret dili mevcuttur. Bu çalışma erişimeaçık Türk işaret dili rakamlarına odaklanmıştır. İşaret dili, toplumun herkesimi tarafından bilinmemektedir. Bu durum, işitme engelli insanlarınbulundukları sosyal ortamlarda iletişim aksaklıklarına neden olmaktadır. İşitmeengelli olmayan ancak işaret dilini bilmeyen bir bireyde aynı problemiyaşamaktadır. Bu çalışmanın amacı, işaret dilini kullanan insanların neanlatmak istediğini derin öğrenme mimarisi üzerinde tespit etmektir. Bu amaçla,işaret dilini rakamlarının, son zamanlarda popülerliği artan siyam sinir ağıile tanımlanmasını bu çalışmada gerçekleştirilmiştir. Siyam sinir ağları,görüntü kümesinde aynı görüntüleri eşleştiren bir derin öğrenme modelidir. Buağları kullanarak Türk işaret dilinde kullanılan rakam görüntülerinitanımlamayı gerçekleştirdik. Elde edilen eşleştirme başarı oranı %98,16’dır.Sonuç olarak, bu çalışma ile Türk işaret dili rakamlarının tanımlanmasındasiyam sinir ağlarının başarılı olduğu görülmüştür.

Graphical abstract, Covid-19 is a disease that affects the upper and lower respiratory tract and has fatal consequences in individuals. Early diagnosis of COVID-19 disease is important. Datasets used in this study were collected from hospitals in Istanbul. The first dataset consists of COVID-19, viral pneumonia, and bacterial pneumonia types. The second dataset consists of the following findings of COVID-19: ground glass opacity, ground glass opacity, and nodule, crazy paving pattern, consolidation, consolidation, and ground glass. The approach suggested in this paper is based on artificial intelligence. The proposed approach consists of three steps. As a first step, preprocessing was applied and, in this step, the Fourier Transform and Gradient-weighted Class Activation Mapping methods were applied to the input images together. In the second step, type-based activation sets were created with three different ResNet models before the Softmax method. In the third step, the best type-based activations were selected among the CNN models using the local interpretable model-agnostic explanations method and re-classified with the Softmax method. An overall accuracy success of 99.15% was achieved with the proposed approach in the dataset containing three types of image sets. In the dataset consisting of COVID-19 findings, an overall accuracy success of 99.62% was achieved with the recommended approach.

White blood cells are cells in the blood and lymph tissue produced by the bone marrow in the human body. White blood cells are an important part of the immune system. The most important task of these cells is to protect the human body against foreign invaders and infectious diseases. When the number of white blood cells in the blood is not enough for the human body, it can cause leukopenia. As a result of this situation, the resistance of the human body against infections and diseases decreases. In this respect, determining the number of these cells in the human body is a specialist task. Detection and treatment of this symptom is a labor-intensive process carried out by specialist doctors and radiologists. Image processing techniques have recently been widely used in biomedical systems for the diagnosis of various diseases. In this study, it is aimed to use image processing techniques to improve the classification performance of deep learning models in white blood cells classification. To perform the classification process more efficiently, the Maximal Information Coefficient and Ridge feature selection methods were used in conjunction with the Convolutional Neural Network models. The Maximal Information Coefficient and Ridge feature selection methods extracted the most relevant features. Afterward, the classification process was realized by using this feature set. In this study, AlexNet, GoogLeNet, and ResNet-50 were used as feature extractor and quadratic discriminant analysis was used as a classifier. As a result, the overall success rate was obtained as 97.95% in the classification of white blood cells. The experimental results showed that the use of the convolutional neural network models with feature selection methods contributed to improving the classification success of white blood cell types.

A proper and certain brain tumor MRI classification has a significant role in current clinical diagnosis, decision making as well as managing the treatment programs. In clinical practice, the examination is performed visually by the specialists, this is a labor-intensive and error-prone process. Therefore, the computer-based systems are in demand so as to carry out objectively this process. In the traditional machine learning approaches, the low-level and high-level handcrafted features used to describe the brain tumor MRI are extracted and classified to overcome the mentioned drawbacks. Considering the recent advances in deep learning, we propose a novel convolutional neural network (CNN) model that is combined with the hypercolumn technique, pretrained AlexNet and VGG-16 networks, recursive feature elimination (RFE), and support vector machine (SVM) in this study. One of the great advantages of the proposed model is that with the help of the hypercolumn technique, it can keep the local discriminative features, which are extracted from the layers located at the different levels of the deep architectures. In addition, the proposed model exploits the generalization abilities of both AlexNet and VGG-16 networks by fusing the deep features achieved from the last fully-connected layers of the networks. Furthermore, the discriminative capacity of the proposed model is enhanced using RFE and thus the most effective deep features are revealed. As a result, the proposed model yielded an accuracy of 96.77% without using any handcrafted feature engine. A fully automated consistent and effective diagnostic model is ensured for the brain tumor MRI classification. Consequently, the proposed model can contribute to realizing a more objective evaluation in the clinics, supporting the decision-making process of the experts, and reducing misdiagnosis rates.

Visualization of the inner ear has been performed using magnetic resonance imaging (MRI) to investigate benign paroxysmal positional vertigo (BPPV). In the patients with BPPV, our recent findings indicate the thickness of some internal auditory canal (IAC) nerves narrower than the thickness of healthy subjects. The thickness of the IAC and its nerves are measured using brain MRI images. The cross sectional area (CSA) of a nerve is assumed as its thickness. Some statistical measurement and a statistical classification are performed on the CSA data to investigate any relation between IAC, the nerves and BPPV.

Recently, the need for automatic identification has caused researchers to focus on biometric identification methods. Palmprint-based biometric identification has several advantages such as user friendliness, low-cost capturing devices, and robustness. In this paper, a method that integrates the discrete cosine transform (DCT) and an autoregressive (AR) signal modeling is proposed for biometric identification. The method provides scale invariance and produces a fixed-length feature vector. In particular, the Burg algorithm is used for the determination of the AR parameters used as a feature vector. Experimental results demonstrate that a small number of the AR parameters that are modeling the DCT coefficients of a palmprint are sufficient to constitute a practically applicable identification system achieving a correct recognition rate of 99.79%. The accuracy of the proposed approach is not overly dependent on the number of training samples, another advantage of the method.

The developments of content based image retrieval (CBIR) systems used for image archiving are continued and one of the important research topics. Although some studies have been presented general image achieving, proposed CBIR systems for archiving of medical images are not very efficient. In presented study, it is examined the retrieval efficiency rate of spatial methods used for feature extraction for medical image retrieval systems. The investigated algorithms in this study depend on gray level co-occurrence matrix (GLCM), gray level run length matrix (GLRLM), and Gabor wavelet accepted as spatial methods. In the experiments, the database is built including hundreds of medical images such as brain, lung, sinus, and bone. The results obtained in this study shows that queries based on statistics obtained from GLCM are satisfied. However, it is observed that Gabor Wavelet has been the most effective and accurate method.

Objectives The aim of the present study was to evaluate the internal auditory canal (IAC) and the nerves inside it to define possible structural differences in cases with subjective tinnitus of unknown origin. Methods Cases applying to the ear, nose and throat department with the complaint of tinnitus with unknown origin and having normal physical examination and test results were included in the study (n=78). Patients admitted to the radiology clinic for routine cranial magnetic resonance imaging (MRI) and whose MRI findings revealed no pathologies were enrolled as the control group (n=79). Data for the control group were obtained from the radiology department and informed consent was obtained from all the patients. Diameters of the IAC and the nerves inside it were measured through enhanced images obtained by routine temporal bone MRIs in all cases. Statistical evaluations were performed using Student t-test and statistical significance was defined as P

Part 4: Pattern Recognition and Image Processing; International audience; Recently, eye states are used as inputs to various applications such as facial expression recognition systems, human-computer interaction and driver fatigue detection systems. Especially with the prominence of human computer interaction, eye state detection has drawn great attention in the past decade. In this study, an eye state detection system based on Circular Hough Transform (CHT) has been offered. Initially, face and eye images are extracted from given gray-level images. After some preprocessing steps, existence of circular iris structure is searched within the extracted eye image using CHT. Existence of circular iris structure is searched within the eye image with the help of circular Hough transform. Eyes are decided as open if iris could identified as a circle.

The wavelet transform (WT) a powerful tool of signal and image processing that have been successfully used in many scientific fields such as signal processing, image compression, computer graphics, and pattern recognition (Daubechies 1990; Lewis and Knowles 1992; Do and Vetterli 2002; Meyer, Averbuch et al. 2002; Heric and Zazula 2007). On contrary the traditional Fourier Transform, the WT is particularly suitable for the applications of nonstationary signals which may instantaneous vary in time (Daubechies 1990; Mallat and Zhang 1993; Akay and Mello 1998). It is crucial to analyze the time-frequency characteristics of the signals which classified as non-stationary or transient signals in order to understand the exact features of such signals (Rioul and Vetterli 1991; Ergen, Tatar et al. 2010). For this reason, firstly, researchers has concentrated on continuous wavelet transform (CWT) that gives more reliable and detailed time-scale representation rather than the classical short time Fourier transform (STFT) giving a time-frequency representation (Jiang 1998; Qian and Chen 1999).