Mekatronik Mühendisliği Bölümü Yayın Koleksiyonu

Bu koleksiyon için kalıcı URI

https://hdl.handle.net/20.500.12885/1686

Listeleniyor 1 - 20 / 59

Multi-image Crowd Counting Using Multi-column Convolutional Neural Network
(Springer, 2022) Kurnaz, Oğuzhan; Hanilçi, Cemal
Crowd density estimation is an important task for security applications. It is a regression problem consisting of feature extraction and estimation steps. In this study, we propose to use a modified version of previously introduced multi-column convolutional neural network (MCNN) approach for estimating crowd density. While in the original MCNN approach the same input image is applied to the each column of the network, we first propose to apply a different version of the same input image to extract a different mapping from each column. Second, original MCNN first generates an estimated density map and then performs crowd counting. Therefore, we adopt it for crowd counting and compare its performance with the proposed method. Regression task is performed by support vector regression (SVR) using feature vectors obtained from MCCNN. 2000 images selected from UCSD pedestrian dataset are used in the experiments. The regions of interest (ROI) are filtered out and the pixel values at the remaining regions are set to zero. In order to prevent distortion caused by camera position, perspective normalization has been applied as a pre-processing step which dramatically improves the performance.
FAULT DIAGNOSIS WITH DEEP LEARNING FOR STANDARD AND ASYMMETRIC INVOLUTE SPUR GEARS
(American Society of Mechanical Engineers (ASME), 2021) Karpat, Fatih; Dirik, Ahmet Emir; Kalay, Onur Can; Yüce, Celalettin; Doğan, Oğuz; Korcuklu, Burak
Gears are critical power transmission elements used in various industries. However, varying working speeds and sudden load changes may cause root cracks, pitting, or missing tooth failures. The asymmetric tooth profile offers higher load-carrying capacity, long life, and the ability to lessen vibration than the standard (symmetric) profile spur gears. Gearbox faults that cannot be detected early may lead the entire system to stop or serious damage to the machine. In this regard, Deep Learning (DL) algorithms have started to be utilized for gear early fault diagnosis. This study aims to determine the root crack for both symmetric and asymmetric involute spur gears with a DL-based approach. To this end, single tooth stiffness of the gears was obtained with ANSYS software for healthy and cracked gears (50-100%), and then the time-varying mesh stiffness (TVMS) was calculated. A six-degrees-of-freedom dynamic model was developed by deriving the equations of motion of a single-stage spur gear mechanism. The vibration responses were collected for the healthy state, 50% and 100% crack degrees for both symmetric and asymmetric tooth profiles. Furthermore, the white Gaussian noise was added to the vibration data to complicate the early crack diagnosis task. The main contribution of this paper is that it adapts the DL-based approaches used for early fault diagnosis in standard profile involute spur gears to the asymmetric tooth concept for the first time. The proposed method can eliminate the need for large amounts of training data from costly physical experiments. Therefore, maintenance strategies can be improved by early crack detection.
Multi-objective optimisation for indentation rate, nugget diameter and tensile load in resistance spot welding using Taguchi-based grey relational analysis
(Inderscience Publishers, 2021) Yüce, Celalettin
In this study, a multi-objective optimisation method based on grey relational analysis with weighted responses is conducted to optimise the RSW parameters of electrode force, welding current, and welding time. Three objectives, such as indentation rate, nugget diameter, and tensile load, are simultaneously optimised. In order to assign the optimum level for each parameter individually, the Taguchi technique was applied. ANOVA results showed that the most influential parameters on indentation rate, nugget diameter, and tensile load are welding current, welding time, and welding current, respectively. In the grey relational analysis, the grey relational grades were obtained using weighted responses. The weight factors for the indentation rate, nugget diameter, and tensile load are 33.24%, 35.67%, and 31.1%, respectively. The optimum parameter combination was obtained as 2,500 N, 9 kA, and 0.5 s. Under these parameter combinations, indentation rate, nugget diameter, and tensile load were 26.2%, 8.34 mm, and 20.04 kN, respectively.
Horizontal attention convolution layer for stereo matching
(Institute of Electrical and Electronics Engineers Inc., 2021) Emlek A.; Peker, Murat
Obtaining a disparity map with stereo matching is one of the most important research topics in areas such as image processing and computer vision. Disparity maps are frequently used by autonomous systems that need depth information of the environment. Recently, high accuracy disparity maps have been obtained with end-to-end deep learning. In this study, a horizontal attention-based convolution layer has been proposed in order to better extract the sequential information in the horizontal plane in the rectified stereo images in methods based on deep learning. The proposed structure has been applied to the DispNetC network, which has been widely used in the literature, and has increased the performance of the network. On the other hand, the proposed method have a very low effect on the network's runtime. The results obtained are shown on the Scene Flow dataset. The codes of the study are available at the following address: https://github.com/aemlek/HADN.
Experimental study on fatigue performance of resistance spot-welded sheet metals
(Springer Science and Business Media Deutschland GmbH, 2021) Ertaş, Ahmet Hanifi; Akbulut M.
Resistance spot welding is used as a reliable joining process in many engineering applications because of its effectiveness, automation capability, and low cost. The spot-welded sheet metals, on the other hand, are prone to mechanical fatigue failure, especially under cyclic loadings. Therefore, understanding and elucidating the fatigue phenomenon of the spot-welded joints are crucial in terms of estimating and preventing undesired failure conditions. In the design phase, there exist a considerable amount of challenges to overcome; one of the most important challenges is to select optimum working conditions. Hence, in this study, the fatigue phenomenon of the spot-welded sheet metals is investigated experimentally, by taking electrode force into consideration. For this purpose, spot-welded modified tensile shear (MTS) test specimens were utilized. A series of fatigue life tests were conducted to examine the influence of electrode force on fatigue life. The results obtained through an optical microscope were presented and interpreted. Experimental data showed that the number of cycles to failure changes depending on the spot-generating schemes in terms of electrode force and welding schedules. Through the investigation of an optical micrograph of partially failed spot-welded MTS specimens for different groups of spot welds created under different electrode force effects, it is seen that the fatigue failure is dominated by the through-thickness cracking. Comparing both crack formation and also fatigue lives of different groups of spot-welded MTS specimens, it is shown that the electrode force and accordingly thermal interaction play an important role in the fatigue strength of the spot-welded specimens.
Emotion recognition using time-frequency ridges of EEG signals based on multivariate synchrosqueezing transform
(De Gruyter Open Ltd, 2021) Mert, Ahmet; Celik H.H.
The feasibility of using time-frequency (TF) ridges estimation is investigated on multi-channel electroencephalogram (EEG) signals for emotional recognition. Without decreasing accuracy rate of the valence/arousal recognition, the informative component extraction with low computational cost will be examined using multivariate ridge estimation. The advanced TF representation technique called multivariate synchrosqueezing transform (MSST) is used to obtain well-localized components of multi-channel EEG signals. Maximum-energy components in the 2D TF distribution are determined using TF-ridges estimation to extract instantaneous frequency and instantaneous amplitude, respectively. The statistical values of the estimated ridges are used as a feature vector to the inputs of machine learning algorithms. Thus, component information in multi-channel EEG signals can be captured and compressed into low dimensional space for emotion recognition. Mean and variance values of the five maximum-energy ridges in the MSST based TF distribution are adopted as feature vector. Properties of five TF-ridges in frequency and energy plane (e.g., mean frequency, frequency deviation, mean energy, and energy deviation over time) are computed to obtain 20-dimensional feature space. The proposed method is performed on the DEAP emotional EEG recordings for benchmarking, and the recognition rates are yielded up to 71.55, and 70.02% for high/low arousal, and high/low valence, respectively.
Convolutional neural networks based rolling bearing fault classification under variable operating conditions
(Institute of Electrical and Electronics Engineers Inc., 2021) Karpat F.; Kalay O.C.; Dirik A.E.; Dogan O.; Korcuklu B.; Yüce, Celalettin
Rolling bearings are key machine elements used in various fields such as automotive, machinery, aviation, and wind turbines. Over time, faults may occur in bearings due to variable operating speeds and loads, contamination, etc., and this may cause a severe reduction in performance. In the future, an undetected bearing fault can lead to a fatal breakdown and substantial economic losses or even human casualties. Thus, bearing early fault diagnosis emerges as a critical and up-to-date topic. It is possible to obtain vibration, acoustic, motor current, etc., data that contain crucial diagnostics information regarding the health conditions of mechanical systems with various sensor technologies. With the era of big data, artificial intelligence (AI) algorithms have started to be utilized frequently in industrial applications. In this regard, convolutional neural networks (CNN) are increasingly popular with their capability to capture fault information without expert knowledge. This paper deals with a bearing fault diagnosis method based on one-dimensional convolutional neural networks (1D CNN) using vibration data. A multi-class classification problem was solved by examining different operating conditions for three health classes. Therefore, healthy state, inner raceway, and outer raceway faults were detected under variable operating speeds (900 and 1500 rpm) and loads (0.1 and 0.7 Nm). The effectiveness of the proposed 1D CNN method was evaluated with the Paderborn University (PU) dataset. As a result, rolling bearing early fault diagnosis was performed with an accuracy of 93.97%. It was observed that the proposed method was suitable for bearing fault diagnosis and can be utilized to optimize the rotary machinery maintenance costs by early fault detection.
A comparative 3d finite element computational study of stress distribution and stress transfer in small-diameter conical dental implants
(Strojarski Facultet, 2021) Kalay O.C.; Karaman H.; Karpat F.; Doğan O.; Yüce, Celalettin
The implant design is one of the main factors in implant stability because it affects the contact area between the bone and the implant surface and the stress-strain distribution at the bone-implant interface. In this study, the effect of different groove geometries on stress-strain distributions in small-diameter conical implants is investigated using the finite element method (FEM). Four different thread models (rectangular, buttressed, reverse buttressed, and symmetrical profile) are created by changing the groove geometry on the one-piece implants, and the obtained results are compared. The stress shielding effect is investigated through the dimensionless numbers that characterize the load-sharing between the bone-implant. It is determined that the lowest stress distribution is observed with rectangular profiled groove geometry. Besides, it is obtained that the buttressed groove geometry minimizes the stress effects transmitted to the periphery of the implant. The symmetrical profiles had better performance than rectangular profiles in stress transfer.
VİDEOLARIN ENF TABANLI ADLİ KANIT ANALİZİNE IŞIK KAYNAĞI ETKİSİ
(2017) Vatansever, Saffet; Dirik, Ahmet Emir
Video dosyalarının Elektrik Şebeke Frekansı (Electric Network Frequency - ENF) temelli adli kanıt analizi tekniği, çoklu ortam dosyalarının kayıt zamanını doğrulamada ve dosyalarda yapılan sahteciliği tespit etmede son yıllarda önerilmiş en önemli araçlardan biridir. ENF, şebekede üretilen toplam gücün tüketilen toplam güce göre artıp azalmasına bağlı olarak nominal değer (Avrupa"da 50 Hz) etrafında sürekli salınımlar yapar. Bu salınımlar aynı şebeke üzerindeki her noktada aynıdır. Elektrik şebekesinden beslenen bir ışık kaynağının yaymış olduğu aydınlatma şiddeti elektrik şebeke frekansına bağlı olarak insan gözünün fark edemeyeceği anlık değişkenlikler gösterir. Işık şiddetindeki bu değişimler, video kameralar tarafından yakalanabilmektedir. Çekilen videolardaki tüm resim çerçeveleri boyunca değişmeyen içerik analiz edilerek aydınlatma şiddetinin değişim hızı, dolayısıyla elektrik şebeke frekansı kestirilebilir. Videolardan kestirimi yapılan ENF sinyalinin, elektrik şebekesinden doğrudan elde edilen referans ENF sinyali ile benzerlikleri hesaplanarak dosya kayıt zamanı bilgisine ulaşılabilir. Bu çalışmada, şebeke elektriğine bağlı ışık kaynağı türünün, CCD sensörlü kamera ile çekilmiş videolardan kestirilen ENF sinyali kalitesinde ne derece etkili olduğu incelenmiştir. Işık kaynağı türüne göre, çeşitli uzunluktaki videolarda ENF temelli kayıt zamanı doğrulama performansı analiz edilmiştir.
Değişken Kuvvetli EMG Sinyallerinin Çok Değişkenli Görgül Kip Ayrışımı ile Analizi ve Sınıflandırılması
(2020) Onay, Fatih; Mert, Ahmet
Elektromiyografi (EMG) sinyalleri, insan-makine etkileşimli akıllı el protezlerinin kontrolünde önemli bir rol oynamaktadır. Kas aktivesinin bir sonucu olarak ortaya çıkan EMG sinyalleri, yapılan aktiviteye dair özel bilgileri kendi içerisinde ihtiva etmektedir. Dolayısıyla akıllı el protezlerinin işlevselliğinin arttırılması, kas bölgesinden toplanan EMG sinyalinin doğru bir şekilde analiz edilip yorumlanmasına önemli ölçüde bağlıdır. Bu konsepte uygun olarak, akıllı el protezi hareketlerinin karar verme sürecinde, EMG sinyallerinin güvenilir bir şekilde kullanılabilmesi için, var olan yöntemlerin geliştirilmesi ya da bu yöntemlere üstünlük sağlayacak yeni yöntemler önerilmesi gerekmektedir. Bu çalışma kapsamında, çok kanallı EMG sinyallerinin analizinin geliştirilmesi amacıyla, çok değişkenli görgül kip ayrışımı (ÇDGKA) tabanlı öznitelik çıkarma yöntemi, geleneksel metotlara alternatif olarak sunulmuştur. Sinyali adaptif olarak salınım modlarına ayıran ÇDGKA yöntemi kullanılarak, EMG sinyalinden daha anlamlı bilgi edinilmesi amaçlanmıştır. ÇDGKA tabanlı özniteliklerin farklı el ve parmak hareketlerini ayırt etme performansı ve farklı kuvvet seviyelerine karşı gösterdiği performans incelenmiştir. Bu amaçla ampute katılımcıların artık uzuvlarından toplanan düşük, orta ve yüksek kuvvet seviyelerine ait EMG sinyalleri üzerinde ÇDGKA yöntemi uygulanarak özgül kip fonksiyonları (ÖKF) elde edilmiştir. Elde edilen ÖKF’lerden çıkarılan öznitelikler kullanılarak altı farklı el ve parmak hareketi, en yakın komşu (k-NN), doğrusal ayrım analizi (LDA) ve destek vektör makinesi (SVM) sınıflandırıcıları kullanılarak sınıflandırılmıştır. Aynı kuvvet seviyesinde eğitilip test edilerek (Senaryo 1) ve tüm kuvvet seviyelerinde eğitilip tek bir kuvvet seviyesinde test edilerek (Senaryo 2) gerçekleştirilen sınıflandırmalar neticesinde, önerilen ÇDGKA tabanlı özniteliklerin ham sinyal tabanlı özniteliklere göre, senaryo 1 için %10 - %15, senaryo 2 için %18’e kadar üstünlük sağladığı belirlenmiştir.
A new high-performance current-mode fuzzy membership function circuit and its application
(2015) Temel, Turgay
A new current-mode fuzzy-membership function circuit is proposed. The circuit has a very compact and simple architecture based on the simultaneous use of winner- and loser-take-all topologies. The circuit has also the advantage of easily adjusting generated membership function characteristics such as center or mean and width with a straightforward application of respective input currents. It is shown that the new circuit outperforms previous counterparts in terms of speed, power consumption, layout area, and robustness to variations in design parameters and errors. As an application, a fuzzy-classiﬁer is designed with a new membership circuit as a seven fuzzy level controller.
Real-time Implementation of Image Based PLC Control for a Robotic Platform
(2019) Ayten, K. K.; Kurnaz, Oğuzhan
In this study, image based real-time control of a linear robotic platform was performed. This robotic platform is used to determine the location of the mushroom and to direct the linear platform to the detected position in real time with PLC control. Haar-Cascade classifier was used to detect mushroom position and Visual Studio C # .NET platform was used to test the Cascade classifier and write other evaluation codes. One of the most important outputs of this work is to determine the actual position in the global coordinate from the pixel-based location of the object in the image using an ordinary USB camera or built-in camera. Calibration technique was used for this determination.
Türkçe için ardışık şartlı rastgele alanlarla bağlılık ayrıştırma
(2017) Bilgin, Metin; Amasyalı, Mehmet Fatih
Sekans etiketleme bir giriş dizisine karşılık bir çıkış dizisinin üretimidir. Giriş ve çıkış dizisinin içeriklerinegöre doğal dil işlemenin birçok konusu (varlık isim tanıma, makine çevirisi, morfolojik analiz, cümleleri öğelerine ayırma vb.) sekans etiketleme olarak tanımlanabilir. Bağlılık ayrıştırması, bir cümle içerisindekisözcükler arasındaki ilişkilerin ve ilişki türlerinin belirlenmesidir ve bir cümlenin anlamsal analizininyapılabilmesi için şarttır. Bağlılık ayrıştırması sekans etiketleme problemi olarak tanımlandığında iki çıkışdizisinin (ilişki türü, ilişkili kelime) birden üretilmesi gerekmektedir. Bizim önerimiz, özellikle Sekansetiketleme problemlerinin çözümünde sıklıkla kullanılan Şartlı Rastgele Alanların bağlılık ayrıştırmasıproblemi içinde kullanılabilir olduğudur. Ancak Şartlı Rastgele Alanlar tek çıkış üreten bir yöntemdir. Buzorluğu aşabilmek için iki çıkışlı (Bağlılık Türü ve Bağlanılan Kelime) bir problem olan Bağlılık Ayrıştırması iki parçaya bölünerek çözülmüştür. Ardından elde edilen sonuçlar birleştirilerek sistemin çıktısıolarak verilmiştir. Gerçekleştirilen bu çalışma ile Türkçe için en yüksek bağlılık ayrıştırması sonuçlarınaulaşılmıştır.
İnsan-robot etkileşiminin biyomimetik yaklaşımla sağlanması
(2019) Gelen, Gökhan; Özcan, Sinan
Bu çalışmada, insan kol ve el hareketlerinin taklit edilmesiyle insanrobot etkileşimini sağlayan biyomimetik bir yaklaşım sunulmuştur. İnsan kol hareketleriyle robotun aynı doğrultuda hareket etmesi sağlanmış ve el hareketleri ile de robot tutucusunun kontrolü sağlanmıştır. Robot hareketi için; ilk olarak insan elinin, bel hizasında orijin noktası olarak belirlenen noktaya olan konumunu verecek kinematik model oluşturulmuştur. Modellemede, insan kolu, ön kol, pazı ve omuz olmak üzere üç ayrı uzuv olarak incelenmiştir. Omuza, pazıya ve ön kola yerleştirilen algılayıcılar ile dönüş açısı bilgileri elde edilmiş ve uzuv uzunlukları ile birlikte matematiksel modelde kullanılmıştır. Bu hesaplamalarda rotasyon kinematiği ve hareket kinematiği matrisleri kullanılmıştır. Tutucu kontrolü için ise bünyesinde EMG sensörleri bulunduran MYO kol bandı kullanılmıştır. Bu kol bandı üzerindeki EMG sensörleri ile kol kaslarından parmak hareketleri algılatılmıştır ve bu hareketler doğrultusunda pnömatik tutucu kontrol edilmiştir. Uygulamalarda 6-eksen robot kolu kullanılmıştır. Hesaplanan konum verileri ve tutucu bilgisi ethernet üzerinden TCP/IP protokolü ile robot denetleyicisine aktarılmaktadır. Robotun hesaplanan konuma gitmesini ve tutucu kontrolünü sağlayan kod oluşturularak robota aktarılmıştır. Yapılan testlerde, endüstriyel robotun insan kol ve el hareketleri ile başarılı biçimde kontrol edildiği gözlemlenmiştir.
Real-time vision-based grasping randomly placed object by low-cost robotic arm using surf algorithm
(IOP Publishing Ltd, 2020) Beyhan, Ayberk; Adar, Nurettin Gökhan
Vision-Based Manipulation is popular and still have open issues in robotics. The camera is a very important part of this method to obtain desired data with image processing techniques. In this study, the Dynamic position-based look and move method was selected to control the 4 DOF robotic arm. For this method, the Kinect camera was used for image processing. Kinect is a special camera which consists of both RGB and infrared camera. SURF algorithm was selected to detect a target object from the target scene using Kinect RGB camera. 3-D target object localization was calculated using Kinect infrared camera with the point cloud. The obtained target object's location is according to the camera and transformed according to the robotic arm base. Using inverse kinematics, desired joint angles were calculated according to the object position. Therefore, the robot is provided to make the desired motion and grasping the object by using gripper. Real-time implementation of the proposed method is carried out using Matlab-Simulink. © 2020 Institute of Physics Publishing. All rights reserved.
Mechatronic System Design of A Smart Mobile Warehouse Robot for Automated Storage/Retrieval Systems
(Institute of Electrical and Electronics Engineers Inc., 2020) Ozbaran, C.; Dilibal, S.; Sungur, Görkem
Smart mobile warehouse robots are widely used in varied industries for increasing the digitalization of the storage and retrieval systems and decreasing the number of workers in warehouses. In this study, the mechatronic system design of a smart mobile warehouse robot is systematically built for automated storage/retrieval system (AS/RS). Each subsystem of the mechatronics design is shaped based on the storage and retrieval tasks. The developed smart mobile warehouse robotic system consists of mechanical, electric-electronic and control subsystems. The mechanical subsystem which consists of telescopic mechanism, scissor lift mechanism, mechanical chassis, and robotic gripper system is integrated to the electric-electronic and control subsystems. Furthermore, the drive system parameters are calculated analytically after creating the motion equations and state-flow diagram for the control subsystem. Structural analyses are conducted using computer-aided simulation programs. Additionally, the robot motion function tests have been explained for the developed smart mobile warehouse robot. © 2020 IEEE.
Effects of rim thickness and drive side pressure angle on gear tooth root stress and fatigue crack propagation life
(Elsevier Ltd, 2021) Doğan, O.; Yüce, Celalettin; Karpat, F.
Gears are the most significant machine elements in power transmission systems. They are used in almost every area of the industry, such as small watches to wind turbines. During the power transmission, gears are subjected to high loads, even unstable conditions, high impact force can be seen. Due to these unexpected conditions, cracks can be seen on the gear surfaces. Moreover, these cracks can propagate, and tooth or body failures can be seen. The fatigue propagation life is related to the gear tooth root stress. If the root stresses decrease, the fatigue life of the gears will increase. In this study, standard and non-standard (asymmetric) gear geometries are formed for four different rim thicknesses and four different pressure angles to examine fatigue crack propagation life. Moreover, the effects of the rim thickness and drive side pressure angle on the root stress are investigated. The static stress analyses are carried out to determine the starting points of the cracks, and the maximum point of the stress is defined as the starting point of the cracks. Fatigue crack propagation analyzes are performed for gears whose crack starting points are determined. The static stress analyses are conducted in ANSYS Workbench; similarly, the fatigue propagation analysis is performed in ANSYS smart crack growth. In this way, the directions of the cracks are determined for different rim thicknesses and drive side pressure angles. Besides, the number of cycles and da/dN graphs is obtained for all cases depending on crack propagation. As a result of the study, maximum stress values were decreased by 66%. The fatigue propagation life was increased approximately fifteen times by using the maximum drive side pressure angle and optimum rim thickness. © 2021 Elsevier Ltd
Effects of drive side pressure angle on gear fatigue crack propagation life for spur gears with symmetric and asymmetric teeth
(American Society of Mechanical Engineers (ASME), 2019) Karpat, F.; Dogan, O.; Yilmaz, T.; Yüce, Celalettin; Kalay, O.C.; Karpat, E.; Kopmaz, O.
Today gears are one of the most crucial machine elements in the industry. They are used in every area of the industry. Due to the high performances of the gears, they are also used in aerospace and wind applications. In these areas due to the high torques, unstable conditions, high impact forces, etc. cracks can be seen on the gear surface. During the service life, these cracks can be propagated and gear damages can be seen due to the initial cracks. The aim of this study is to increase the fatigue crack propagation life of the spur gears by using asymmetric tooth profile. Nowadays asymmetric gears have a very important and huge usage area in the industry. In this study, the effects of drive side pressure angle on the fatigue crack propagation life are studied by using the finite element method. The initial starting points of the cracks are defined by static stress analysis. The starting angles of the cracks are defined constant at 45°. The crack propagation analyses are performed in ANSYS SMART Crack-Growth module by using Paris Law. Four different drive side pressure angles (20°-20°, 20°-25°, 20°-30° and 20°-35°) are investigated in this study. As a result of the study the fatigue crack propagation life of the gears is increased dramatically when the drive side pressure angle increase. This results show that the asymmetric tooth profile not only decrease the bending stress but also increase the fatigue crack propagation life strongly. Copyright © 2019 ASME.
Design of an Event Based Multi Processors Communication System
(Institute of Electrical and Electronics Engineers Inc., 2017) Kart, M.; Gul, E.; Türe, Murat
This paper presents an event based model multi-processor system that can handle events from hundreds ofterminals. The overall system acts as a single system as if it iscontrolled by a single processor. A communication system wasimplemented using the proposed model. Every node in the systemruns the same program and in order to add new processing nodesinto the system only the configuration files have to be changed. © 2017 IEEE.
Multi-Image Crowd Density Estimation using Multi Column Deep Neural Network
(Institute of Electrical and Electronics Engineers Inc., 2020) Kurnaz, Oğuzhan; Hanilçi, Cemal
Crowd density estimation is a challenging problem for security applications which is a regression task consisting of feature extraction and regression steps. In this paper, multicolumn deep neural networks (MDNN) is proposed for crowd density estimation. Regression task is performed using the feature vector obtained from MDNN. 2000 images selected from the UCSD pedestrian dataset are used in the experiments. In the images, the region of interest (ROI) is filtered and the remaining parts are removed. In order to avoid distortions due to camera position, perspective normalization is applied as a pre-processing step which yields considerable performance improvement. © 2020 IEEE.

Güncel Gönderiler