Browsing by Author "Sumer, Emre"

Now showing 1 - 20 of 21

Automated Extraction of Photorealistic Facade Textures from Single Ground-Level Building Images
(2014) Sumer, Emre; Turker, Mustafa; AGA-5711-2022
An integrated approach is presented for the automatic extraction of photorealistic facade textures from single street-level building images. The initial facade texture is extracted using Watershed segmentation. The seed pixels (markers) to trigger the segmentation are located automatically both for the foreground (facade) and the background regions, and the segmentation is carried out repetitively until the facade texture is extracted. The extracted facade image is geometrically rectified using a developed automatic technique based on Hough transformation and interest point detection. The occluded areas on facade textures are restored by employing an image matching-based procedure. The approach was tested on two different datasets captured from the residential areas of Ankara, the capital of Turkey. The datasets contain a total of 40 building facade images that were taken from the street-level. The results indicate that the facade textures are extracted adequately. For facade image extraction, an average quantitative accuracy of 83% was achieved. For rectification, 24 out of 40 buildings provided the positional error under 10 pixels at 95% confidence level. The subjective assessment of facade restoration yielded the mean rating value of 2.46 for the datasets used, in which the rating values are ranked between 1 for "Excellent" and 6 for "Unusable".
Automatic Near-Photorealistic 3-D Modelling and Texture Mapping for Rectilinear Buildings
(2017) Sumer, Emre; Turker, Mustafa; AGA-5711-2022
Three-dimensional (3-D) representations of urban regions have gained much attention because of recent developments in remote sensing and computer graphics technologies. In particular, textured 3-D building reconstruction for a variety of applications has been a popular research topic in recent years. In this study, we present the reconstruction of 3-D building models along with texture selection and mapping. Extracted two-dimensional building patches and normalized digital surface model (nDSM) data are used to generate the 3-D models. To build near-photorealistic 3-D models, the acquired geo-referenced facade textures are associated with the corresponding building facades using an automated GPS-assisted approach. On the other hand, the modelling and texture mapping of the roof structures were carried out manually. The study area is composed of eight housing estates (blocks), where a total of 110 buildings were analysed. The whole study area was modelled, with facade textures, in less than 1min of processor running time with an acceptable level of accuracy. The texture mapping was carried out using MATLAB's Virtual Reality Toolbox.
CNN-Based Severity Prediction Of Neurodegenerative Diseases Using Gait Data
(2022) Erdas, Cagatay Berke; Sumer, Emre; Kibaroglu, Seda; https://orcid.org/0000-0003-3467-9923; 35111334; AGA-5711-2022
Neurodegenerative diseases occur because of degeneration in brain cells but can manifest as impairment of motor functions. One of the side effects of this impairment is an abnormality in walking. With the development of sensor technologies and artificial intelligence applications in recent years, the disease severity of patients can be estimated using their gait data. In this way, decision support applications for grading the severity of the disease that the patient suffers in the clinic can be developed. Thus, patients can have treatment methods more suitable for the severity of the disease. The presented research proposes a deep learning-based approach using gait data represented by a Quick Response code to develop an effective and reliable disease severity grading system for neurodegenerative diseases such as amyotrophic lateral sclerosis, Huntington's disease, and Parkinson's disease. The two-dimensional Quick Response data set was created by converting each one-dimensional gait data of the subjects with a novel representation approach to a Quick Response code. This data set was regressed with the convolutional neural network deep learning method, and a solution was sought for the problem of grading disease severity. Further, to demonstrate the success of the results obtained with the novel approach, native machine learning approaches such as Multilayer Perceptron, Random Forest, Extremely Randomized Trees, and K-Nearest Neighbours, and ensemble machine learning methods, such as voting and stacking, were applied on one-dimensional data. Finally, the results obtained on the prediction of disease severity by testing one-dimensional gait data with a convolutional neural network architecture that operates on one-dimensional data were included. The results showed that, in most cases, the two-dimensional convolutional neural network approach performed the best among all methods.
Comparison of Different Machine Learning Approaches to Detect Femoral Neck Fractures in X-Ray Images
(2021) Acici, Koray; Sumer, Emre; Beyaz, Salih; 0000-0002-3821-6419; 0000-0001-8502-9184; 0000-0002-5788-5116; AGA-5711-2022; K-8820-2019
Femoral neck fractures are a serious health problem, especially in the elderly population. Misdiagnosis leads to improper treatment and adversely affects the quality of life of the patients. On the other hand, when looking from the perspective of orthopedic surgeons, their workload increases during the pandemic, and the rates of correct diagnosis may decrease with fatigue. Therefore, it becomes essential to help healthcare professionals diagnose correctly and facilitate treatment planning. The main purpose of this study is to develop a framework to detect fractured femoral necks in PXRs (Pelvic X-ray, Pelvic Radiographs) while also researching how different machine learning approaches affect different data distributions. Conventional, LBP (Local Binary Patterns), and HOG (Histogram of Gradients) features were extracted manually from gray-level images to feed the canonical machine learning classifiers. Gray-level and three-channel images were used as inputs to extract the features automatically by CNNs (Convolutional Neural Network). LSTMs (Long Short-Term Memory) and BILSTMs (Bidirectional Long Short-Term Memory) were fed by automatically extracted features. Metaheuristic optimization algorithms, GA (Genetic Algorithm) and PSO (Particle Swarm Optimization), were utilized to optimize hyper-parameters such as the number of the feature maps and the size of the filters in the convolutional layers of the CNN architecture. The majority voting was applied to the results of the different classifiers. For the imbalanced dataset, the best performance was achieved by the 2-layer LSTM architecture that used features extracted from the fifth max-pooling layer of the CNN architecture optimized by GA. For the balanced dataset, the best performance was obtained by the CNN architecture optimized by PSO in terms of the Kappa evaluation metric. Although metaheuristic optimization algorithms such as GA and PSO do not guarantee the optimal solution, they can improve the performance on a not extremely imbalanced dataset especially in terms of sensitivity and Kappa evaluation metrics. On the other hand, for a balanced dataset, more reliable results can be obtained without using metaheuristic optimization algorithms but including them can result in an acceptable agreement in terms of the Kappa metric.
Context-Sensitive Model Learning for Lung Nodule Detection
(2016) Ogul, B. Buket; Ogul, Hasan; Sumer, Emre; AGA-5711-2022
Nodule detection in chest radiographs is a main component of current Computer Aided Diagnosis (CAD) systems. The problem is usually approached as a supervised classification task of candidate nodule segments. To this end, a discriminative model is learnt from predefined set of features. A key concern with this approach is the fact that some normal tissues are also imaged and these regions can overlap with the lung tissue as to hide the nodules. These overlaps may reduce the discriminative ability of extracted features and increase the number of false positives accordingly. In this study, we offer to learn distinct models for bone and normal tissue regions following to the segmentation of ribs, which are often the major reason for false positives. Thus, the nodule candidates in bone and normal tissue regions can be assessed in context-sensitive way. The experiments on a common benchmark set determine that the proposed approach can significantly recue the false positives while preserving the sensitivity of detections.
Creation and Solution of Image Processing Based CAPTCHA Test
(2018) Topaloglu, Onur; Bulut, Tunahan; Ates, Umut; Timur, Kenan; Sumer, Emre; AGA-5711-2022
A Deep Learning Based Approach to Detect Neurodegenerative Diseases
(2020) Erdas, Cagatay Berke; Sumer, Emre; AGA-5711-2022
Studies conducted by the World Health Organization (WHO) show that more than a billion people worldwide suffer from neurological disorders and the lack of effective diagnostic procedures affects treatment. Characterizing specific motor symptoms to facilitate their diagnosis can he useful in monitoring disease progression and effective treatment planning. Classification of highly prevalent neurodegenerative diseases (NDD) such as Parkinson's disease (PH), Amyotrophic Lateral Sclerosis (ALS), and Huntington's disease (HH) is of clinical importance, One of the methods used in the literature to detect these neurodegenerative diseases is gait analysis-based classification. in this study, 12 different features fed a unidimensional Convolutional Neural Network (CNN) deep learning algorithm-based model, and aims to detect PD, HD, and ALS diseases was studied. The unidimensional CNN deep learning model fed with 12 features achieved 78,92%, 84,39% and 92,09% classification accuracy for control against HH, control. against PH, and control detection problems against ALS. Again, the relevant classifier produced 84,75% accuracy with the approach developed to separate all neurodegenerative disease specimens (NDD) under a single label as a class, and to distinguish these diseases against the current control.
Eliminating Rib Shadows in Chest Radiographic Images Providing Diagnostic Assistance
(2016) Ogul, Hasan; Ogul, B. Buket; Agildere, A. Muhtesem; Bayrak, Tuncay; Sumer, Emre; https://orcid.org/0000-0003-4223-7017; https://orcid.org/0000-0001-6826-4350; 26775736; AAB-5802-2020; U-4603-2019
A major difficulty with chest radiographic analysis is the invisibility of abnormalities caused by the superimposition of normal anatomical structures, such as ribs, over the main tissue to be examined. Suppressing the ribs with no information loss about the original tissue would therefore be helpful during manual identification or computer-aided detection of nodules on a chest radiographic image. In this study, we introduce a two-step algorithm for eliminating rib shadows in chest radiographic images. The algorithm first delineates the ribs using a novel hybrid self-template approach and then suppresses these delineated ribs using an unsupervised regression model that takes into account the change in proximal thickness (depth) of bone in the vertical axis. The performance of the system is evaluated using a benchmark set of real chest radiographic images. The experimental results determine that proposed method for rib delineation can provide higher accuracy than existing methods. The knowledge of rib delineation can remarkably improve the nodule detection performance of a current computer-aided diagnosis (CAD) system. It is also shown that the rib suppression algorithm can increase the nodule visibility by eliminating rib shadows while mostly preserving the nodule intensity. (C) 2015 Elsevier Ireland Ltd. All rights reserved.
Encryption and multi-share-based seganography methods on images with low spectral resolution
(Başkent Üniversitesi Fen Bilimler Enstitüsü, 2023) Çiftçi, Efe; Sumer, Emre
Steganography is the name given to secret communication methods that third parties cannot detect. This secret communication is performed by hiding the secret information to be transmitted on a carrier medium so that the carrier does not raise any suspicions. Steganography science, of which many examples can be presented from the past to the present, has gained new application areas with the development of digital technologies. This thesis aims to develop new steganography methods that hide secret messages in plain text format on binary images, which have a lower spectral resolution when compared to color or grayscale images, used in digital devices as carriers. It has been observed that all implemented methods can successfully hide considerable lengths of plaintext payloads on binary images generated by both thresholding and halftoning methods, and this finding has been reinforced with conducted objective and subjective evaluations. Steganografi, üçüncü şahıslar tarafından tespit edilmeyecek şekilde gizli iletişim kurma yöntemlerine verilen isimdir. Bu gizli iletişim, iletilmek istenen gizli bilginin şüphe uyandırmayacak bir şekilde bir taşıyıcı ortamın üzerine gizlenmesiyle gerçekleşir. Geçmişten günümüze bir çok örneği sunulabilen steganografi bilimi, dijital teknolojilerin gelişmesiyle yeni uygulama alanları kazanmıştır. Bu tezin amacı, düz metin biçimindeki gizli mesajları, taşıyıcı olarak dijital cihazlarda kullanılan renkli veya gri tonlu görüntülere kıyasla daha düşük spektral çözünürlüğe sahip ikili görüntüler üzerine gizleyecek olan yeni steganografi yöntemlerinin geliştirilmesidir. Geliştirilen tüm yöntemlerin hem eşikleme, hem de yarıtonlama yöntemleriyle üretilen ikili görüntülere büyük uzunluklarda düz metin türünde veriyi başarıyla gizleyebildikleri görülmüş ve yapılan objektif ile subjektif değerlendirmelerle de bu bulgu pekiştirilmiştir.
An Eye Controlled Wearable Communication and Control System for Als Patients: Smarteyes
(2017) Sumer, Emre; Uslu, I. Baran; Turker, Mustafa; 0000-0001-5604-0472; AGA-5711-2022; AAR-1071-2020; G-6129-2013
ALS (Amyotrophic Lateral Sclerosis) is a progressive neurodegenerative disease that involves the malfunctioning of motor neurons. The ability of the brain to initiate and control muscle movement is lost subsequent to death of motor neurons. People with ALS present the greatest challenge regarding communication issues. Besides, caring for a loved one with ALS is not an easy task. In this study, we developed an eye-controlled wearable system called "SmartEyes" which improves the life qualities of ALS patients and their caregivers by offering two important skills. The first skill is communicating through predefined voice messages generated by a computer and the second one is controlling several peripherals located in the patient's environment. The developed system is novel in that; the patients can easily vocalize their needs and requests with a few sequential eye movements. Moreover, they can control several household items including desk lamp, rolling curtain, television and air conditioner in the same way. The preliminary experiments showed that the performance of the system is satisfactory. The accuracy of the system commands based on pupil gaze direction was tested on several users and about an accuracy of 89% was achieved. It is believed that the developed system has attracted the patients' and their caregivers' interest very much and this is the main motivation in improving our system.
An Eye-Controlled Wearable Communication and Control System for Als Patients: Smarteyes
(2017) Sumer, Emre; Uslu, I. Baran; Turker, Mustafa; AGA-5711-2022
ALS (Amyotrophic Lateral Sclerosis) is a progressive neurodegenerative disease that involves the malfunctioning of motor neurons. The ability of the brain to initiate and control muscle movement is lost subsequent to death of motor neurons. People with ALS present the greatest challenge regarding communication issues. Besides, caring for a loved one with ALS is not an easy task. In this study, we developed an eye-controlled wearable system called "SmartEyes" which improves the life qualities of ALS patients and their caregivers by offering two important skills. The first skill is communicating through predefined voice messages generated by a computer and the second one is controlling several peripherals located in the patient's environment. The developed system is novel in that; the patients can easily vocalize their needs and requests with a few sequential eye movements. Moreover, they can control several household items including desk lamp, rolling curtain, television and air conditioner in the same way. The preliminary experiments showed that the performance of the system is satisfactory. The accuracy of the system commands based on pupil gaze direction was tested on several users and about an accuracy of 89% was achieved. It is believed that the developed system has attracted the patients' and their caregivers' interest very much and this is the main motivation in improving our system.
Femoral neck fracture detection in X-ray images using deep learning and genetic algorithm approaches
(2020) Beyaz, Salih; Acici, Koray; Sumer, Emre; 0000-0002-5788-5116; 32584712; K-8820-2019
Objectives: This study aims to detect frontal pelvic radiograph femoral neck fracture using deep learning techniques. Patients and methods: This retrospective study was conducted between January 2013 and January 2018. A total of 234 frontal pelvic X-ray images collected from 65 patients (32 males, 33 females; mean age 74.9 years; range, 33 to 89 years) were augmented to 2106 images to achieve a satisfactory dataset. A total of 1,341 images were fractured femoral necks while 765 were non-fractured ones. The proposed convolutional neural network (CNN) architecture contained five blocks, each containing a convolutional layer, batch normalization layer, rectified linear unit, and maximum pooling layer. After the last block, a dropout layer existed with a probability of 0.5. The last three layers of the architecture were a fully connected layer of two classes, a softmax layer and a classification layer that computes cross entropy loss. The training process was terminated after 50 epochs and an Adam Optimizer was used. Learning rate was dropped by a factor of 0.5 on every five epochs. To reduce overfitting, regularization term was added to the weights of the loss function. The training process was repeated for pixel sizes 50x50, 100x100, 200x200, and 400x400. The genetic algorithm (GA) approach was employed to optimize the hyperparameters of the CNN architecture and to minimize the error after testing the model created by the CNN architecture in the training phase. Results: Performance in terms of sensitivity, specificity, accuracy, F1 score, and Cohen's kappa coefficient were evaluated using five-fold cross validation tests. Best performance was obtained when cropped images were rescaled to 50x50 pixels. The kappa metric showed more reliable classifier performance when 50x50 pixels image size was used to feed the CNN. The classifier performance was more reliable according to other image sizes. Sensitivity and specificity rates were computed to be 83% and 73%, respectively. With the inclusion of the GA, this rate increased by 1.6%. The detection rate of fractured bones was found to be 83%. A kappa coefficient of 55% was obtained, indicating an acceptable agreement. Conclusion: This experimental study utilized deep learning techniques in the detection of bone fractures in radiography. Although the dataset was unbalanced, the results can be considered promising. It was observed that use of smaller image size decreases computational cost and provides better results according to evaluation metrics.
Gender Recognition Using Innovative Pattern Recognition Techniques
(2018) Kabasakal, Burak; Sumer, Emre; AGA-5711-2022
The vast number of researchers has been focused on pattern recognition and computer vision fields in parallel with recent technological developments over the last two decades. Some of the topics in these areas are; face detection, face recognition and gender recognition. Mostly because, the studies conducted on these areas use native ways to collect biometric data without causing any inconvenience to the subject with their contactless and free flow nature. In this paper, a new system that provides gender information using facial images is presented. The system consists of two main stages; (i) face detection and (ii) gender recognition. In the first stage, the system focuses on the detection of frontal human faces in digital images. We used a linear classifier combined with Histogram of Oriented Gradients (HOG) feature for face detection. In the second stage, two different classifiers for gender recognition were trained. The first classifier is based on Support Vector Machines (SVM) and the second is based on Convolutional Neural Networks (CNN) which is also known as Deep Learning. We used Local Binary Pattern (LBP) and HOG as features for SVM classifier, and Radial Basis Function (RBP) as its kernel. For the CNN classifier, we used GoogleNet deep neural network architecture and the optimization was performed depending on the parameters. For training of both classifiers, Labeled Faces in the Wild (LFW), IMDB and WIKI data sets were used. In our experiments, we observed that the CNN based classifier surpasses the SVM based one in terms of accuracy.
Hand and Pose-Based Feature Selection for Zero-Shot Sign Language Recognition
(IEEE ACCESS, 2024-08-22) Ozcan, Giray Sercan; Bilge, Yunus Can; Sumer, Emre
Sign language functions as an indispensable interaction method for a certain portion of people in society, offering a unique way of communication. A significant challenge in advancing towards this objective is the difficulty in obtaining suitable training data for each sign in supervised learning. This challenge comes from the complex process of labeling signs and the limited number of skilled people available to do this job. This work introduces a new approach to the problem of Zero-Shot Sign Language Recognition (ZSSLR). We basically utilize and model hand and landmark data streams extracted from the body of the signer. Based on these extracted and modeled features, we employ a data grading approach to facilitate visual embedding with the self-attention mechanism. We utilize textual sign description features along with visual embedding in the Zero-Shot Learning (ZSL) settings. We assess the efficacy of our methodology in two of the suggested ZSL benchmarks.
Monitoring nodule progression in chest X-ray images
(2018) Sumer, Emre; Engin, Muharrem; Agildere, Muhtesem; Ogul, Hasan
Lung nodules are frequently observed in cases of cancer. Nodules can be monitored with technologies such as computed tomography (CT) or magnetic resonance imaging (MRI). However. x-ray imaging is a low-cost method as well as its widespread usage. In this context, monitoring the nodules in short intervals by x-ray imaging gives benefits in many aspects. In this study, a three-stage novel approach is proposed to trace the nodule progressions from the lung x-ray images, automatically. In the first stage, x-ray images of a patient taken at different times must be registered to evaluate the nodule progression. To perform the registration, feature extraction and matching methods are employed, and then the homography matrix is calculated. In the second stage, according to previously known nodule positions, matched nodules are detected on registered images. Mismatched nodules in the first image are considered as lost, while the nodules only found in the second image are evaluated as newly appeared. In the last stage, nodules are considered as closed contours consisting of pixel set where closed contour area is calculated after nodule matching process. In this way, growth and shrink states are determined numerically. To test the proposed approach, a patient data set provided by Baskent University, Department of Radiology is used. The validation of the test results is performed by an expert radiologist According to the results obtained, the presented nodule progression trace system is found promising.
Neurodegenerative disease detection and severity prediction using deep learning approaches
(2021) Erdas, Cagatay Berke; Sumer, Emre; Kibaroglu, Seda; 0000-0002-3964-268X; AAJ-2956-2021; AGA-5711-2022
Neurodegenerative diseases (NDDs) such as amyotrophic lateral sclerosis (ALS), Huntington's disease (HD), and Parkinson's disease (PD) can manifest themselves anatomically by degeneration in the brain as well as motor symptoms. The motor symptoms can affect walking dynamics in a disease-specific fashion; characteristically they disrupt gait. As the severity of the disease increases, walking ability deteriorates. We examined the effect of NDDs such as ALS, HD, and PD on gait and developed a convolutional long short-term memory (ConvLSTM) and threedimensional convolutional learning network (3D CNN)-based approach to detecting neurodegenerative conditions and predicting disease severity.
A novel steganography method for binary and color halftone images
(2022) Ciftci, Efe; Sumer, Emre; 36091978
Performance Analysis of Spatial and Frequency Domain Filtering In High Resolution Images
(2015) Asuroglu, Tunc; Sumer, Emre; 0000-0003-4153-0764; ITV-2441-2023; AGA-5711-2022
Promoting Development Through A Geographic Information System-Based Lodging Property Query System (LPQS) for Antalya, Turkey
(2016) Sumer, Selay Ilgaz; Sumer, Emre; Atasever, Hilal; AGA-5711-2022
Information technology currently plays an important role in many industries and has enabled the development of different sectors and economies. Geographic information system (GIS) is an information technology that triggers improvements in many countries, and this paper presents a method of using GIS in the retrieval of lodging properties. A Lodging Property Query System (LPQS) is a novel system proposed for use by travel agencies to perform spatial queries. The proposed system was tested on a sample dataset that contains lodging properties selected from five different regions located along the shoreline of Antalya, Turkey. The data layers were prepared with the MapInfo software package, and the spatial queries and graphical user interface were developed with the MapXtreme software development kit. This study aims to contribute to the development of the travel agencies by offering useful information that fits customer expectations and needs by means of spatial context.
WEEE estimation and determination of collection points: A case for the Municipality of Cankaya
(2018) Kececi, Baris; Dengiz, Orhan; Dengiz, Berna; Sumer, Emre; Kilic, Aybuke; Ceki, Ece; Inan, Bedia; Cicek, Selda; 0000-0002-2730-5993; F-1639-2011; AAC-4793-2019
Electrical-electronic waste (WEEE) is generated from electric and electronic devices at the end of their lifecycles. When WEEEs are disposed, burned or disassembled for recycling, they are transformed into products containing hazardous substances and harm the environment and human health. Therefore WEEEs should be collected, transported and processed properly during their recycling operations by municipality authority in order to minimize their damage and maximize their economic benefit. In this paper, a study is carried out for the WEEE management of Cankaya Municipality by the cooperation of Environmental Conservation and Control Department of Cankaya. For this purpose first WEEE with a focus on television, refrigerator, washing machine, oven and vacuum cleaner are estimated based on the electronic device usage behavior of residents in Cankaya. Secondly, the WEEE collection points' location problem is solved considering set covering problem.