18 results
Search Results
Now showing 1 - 10 of 18
Article Citation - WoS: 6Classification of Parasite Egg Cells Using Gray Level Cooccurence Matrix and Knn(Scientific Publishers india, 2016) Sengul, GokhanParasite eggs are around 20 to 80 mu m dimensions, and they can be seen under microscopes only and their detection requires visual analyses of microscopic images, which requires human expertise and long analysis time. Besides visual analysis is very error prone to human procedures. In order to automatize this process, a number of studies are proposed in the literature. But there is still a gap between the preferred performance and the reported ones and it is necessary to increase the performance of the automatic parasite egg classification approaches. In this study a learning based statistical pattern recognition approach for parasite egg classification is proposed that will both decrease the time required for the manual classification by an expert and increase the performance of the previously suggested automated parasite egg classification approaches. The proposed method uses Gray-Level Co-occurrence Matrix as the feature extractor, which is a texture based statistical method that can differentiate the parasite egg cells based on their textures, and the k-Nearest Neighbourhood (kNN) classifier for the classification. The proposed method is tested on 14 parasite egg types commonly seen in humans. The results show that proposed method can classify the parasite egg cells with a performance rate of 99%.Conference Object Citation - WoS: 1Gender Prediction by Using Local Binary Pattern and K Nearest Neighbor and Discriminant Analysis Classifications(Ieee, 2016) Camalan, Seda; Çamalan, Seda; Sengul, Gokhan; Şengül, Gökhan; Çamalan, Seda; Şengül, Gökhan; Information Systems Engineering; Computer Engineering; Information Systems Engineering; Computer EngineeringIn this study, gender prediction is investigated for the face images. To extract the features of the images, Local Binary Pattern (LBP) is used with its different parameters. To classify the images male or female, K-Nearest Neighbors (KNN) and Discriminant Analysis (DA) methods are used. Their performances according to the LBP parameters are compared. Also classification methods' parameters are changed and the comparison results are shown. These methods are applied on FERET database with 530 female and 731 male images. To have better performance, the face parts of the images are cropped then feature extraction and classification methods applied on the face part of the images.Article Citation - WoS: 9Citation - Scopus: 9Benchmarking Classification Models for Cell Viability on Novel Cancer Image Datasets(Bentham Science Publ Ltd, 2019) Ozkan, Akin; Isgor, Sultan Belgin; Sengul, Gokhan; Isgor, Yasemin GulgunBackground: Dye-exclusion based cell viability analysis has been broadly used in cell biology including anticancer drug discovery studies. Viability analysis refers to the whole decision making process for the distinction of dead cells from live ones. Basically, cell culture samples are dyed with a special stain called trypan blue, so that the dead cells are selectively colored to darkish. This distinction provides critical information that may be used to expose influences of the studied drug on considering cell culture including cancer. Examiner's experience and tiredness substantially affect the consistency throughout the manual observation of cell viability. The unsteady results of cell viability may end up with biased experimental results accordingly. Therefore, a machine learning based automated decision-making procedure is inevitably needed to improve consistency of the cell viability analysis. Objective: In this study, we investigate various combinations of classifiers and feature extractors (i.e. classification models) to maximize the performance of computer vision-based viability analysis. Method: The classification models are tested on novel hemocytometer image datasets which contain two types of cancer cell images, namely, caucasian promyelocytic leukemia (HL60), and chronic myelogenous leukemia (K562). Results: From the experimental results, k-Nearest Neighbor (KNN) and Random Forest (RF) by combining Local Phase Quantization (LPQ) achieve the lowest misclassification rates that are 0.031 and 0.082, respectively. Conclusion: The experimental results show that KNN and RF with LPQ can be powerful alternatives to the conventional manual cell viability analysis. Also, the collected datasets are released from the "biochem.atilim.edu.tr/datasets/ " web address publically to academic studies.Article Citation - WoS: 8Estimation of Polypropylene Concentration of Modified Bitumen Images by Using K-Nn and Svm Classifiers(Asce-amer Soc Civil Engineers, 2015) Tapkin, Serkan; Sengoz, Burak; Sengul, Gokhan; Topal, Ali; Ozcelik, ErolThe goal of this study is to design an expert system that automatically classifies the microscopic images of polypropylene fiber (PPF) modified bitumen including seven different contents of fibers. Optical microscopy was used to capture the images from thin films of polypropylene fiber modified bitumen samples at a magnification scale of 100 x. A total of 313 images were pre-processed, and features were extracted and selected by the exhaustive search method. The k-nearest neighbor (k-NN) and multiclass support vector machine (SVM) classifiers were applied to quantify the representation capacity. The k-NN and multiclass SVM classifiers reached an accuracy rate of 87% and 86%, respectively. The results suggest that the proposed expert system can successfully estimate the concentration of PPF in bitumen images with good generalization characteristics. (C) 2014 American Society of Civil Engineers.Article Citation - WoS: 47Citation - Scopus: 66Deep Learning Based Fall Detection Using Smartwatches for Healthcare Applications(Elsevier Sci Ltd, 2022) Sengul, Gokhan; Karakaya, Murat; Misra, Sanjay; Abayomi-Alli, Olusola O.; Damasevicius, RobertasWe implement a smart watch-based system to predict fall detection. We differentiate fall detection from four common daily activities: sitting, squatting, running, and walking. Moreover, we separate falling into falling from a chair and falling from a standing position. We develop a mobile application that collects the acceleration and gyroscope sensor data and transfers them to the cloud. In the cloud, we implement a deep learning algorithm to classify the activity according to the given classes. To increase the number of data samples available for training, we use the Bica cubic Hermite interpolation, which allows us to improve the accuracy of the neural network. The 38 statistical data features were calculated using the rolling update approach and used as input to the classifier. For activity classification, we have adopted the bi-directional long short-term memory (BiLSTM) neural network. The results demonstrate that our system can detect falling with an accuracy of 99.59% (using leave-one-activityout cross-validation) and 97.35% (using leave-one-subject-out cross-validation) considering all activities. When considering only binary classification (falling vs. all other activities), perfect accuracy is achieved.Article Citation - WoS: 8Citation - Scopus: 13White Blood Cells Classifications by SURF Image Matching, PCA and Dendrogram(Allied Acad, 2015) Nazlibilek, Sedat; Karacor, Deniz; Erturk, Korhan Levent; Sengul, Gokhan; Ercan, Tuncay; Aliew, Fuad; Department of Mechatronics Engineering; Information Systems Engineering; Computer EngineeringDetermination and classification of white blood cells are very important for diagnosing many diseases. The number of white blood cells and morphological changes or blasts of them provide valuable information for the positive results of the diseases such as Acute Lymphocytic Leucomia (ALL). Recognition and classification of white cells as basophils, lymphocytes, neutrophils, monocytes and eosinophils also give additional information for the diagnosis of many diseases. We are developing an automatic process for counting, size determination and classification of white blood cells. In this paper, we give the results of the classification process for which we experienced a study with hundreds of images of white blood cells. This process will help to diagnose especially ALL disease in a fast and automatic way. Three methods are used for classification of five types of white blood cells. The first one is a new algorithm utilizing image matching for classification that is called the Speed-Up Robust Feature detector (SURF). The second one is the PCA that gives the advantage of dimension reduction. The third is the classification tree called dendrogram following the PCA. Satisfactory results are obtained by two techniques.Article The Effect of Statistically Constrained Minimum Mean Square Estimation Algorithm Which Is Used for Human Head Tissue Conductivity Estimation To Source Localization(Journal Neurological Sciences, 2012) Sengul, Gokhan; Şengül, Gökhan; Baysal, Ugur; Şengül, Gökhan; Computer Engineering; Computer Engineering; Computer EngineeringDetermining the electrical active regions of human brain by using EEG and/or MEG data is known as "EEG/MEG bioelectromagnetic inverse problem" or "source localization". A typical source localization system intakes not only EEG/MEG data but also geometry information of subject/patient, a priori information about the electrically active sources, the number and 3-D positions of measurement electrodes and conductivities/resistivities of the tissues in the head model. In this study we investigated the conductivity estimation performance previously proposed Statistically Constrainted Minimum Mean Square Error Estimation (MiMSEE) algorithm by simulation studies and we also investigated the effect of the estimation to source localization activities. In simulation studies we used a three-layered (composed of scalp, skull and brain regions) realistic head model to estimate 100 different conductivity distributions in vivo. As a result we found that the proposed algorithm estimates the conductivity of scalp with an average error of 23%, the conductivity of skull with an average error of 40% and finally the conductivity of brain with an average error of 17%. In the second part of the study we compared the source localization errors for two cases: one, when the average conductivities of tissues given in the literature are used, and second when the subject-specific conductivity estimation is performed with MiMSEE algorithm. The results showed 10.1 mm localization error is obtained when the average conductivities given in the literature are used and 2.7 mm localization is obtained when subject-specific conductivity estimation is performed with MiMSEE algorithm. The results shows that the localization error is reduced by 73.07% when subject-specific conductivity estimation is performed with MiMSEE algorithm. We conclude that using the conductivities obtained from MiMSEE algorithm reduces the source localization error and we recommend to perform subject-specific conductivity estimation for source localization applications.Article Citation - WoS: 5Citation - Scopus: 5A Study on the Performance of Magnetic Material Identification System by Sift-Brisk and Neural Network Methods(Ieee-inst Electrical Electronics Engineers inc, 2015) Ege, Yavuz; Nazlibilek, Sedat; Kakilli, Adnan; Citak, Hakan; Kalender, Osman; Karacor, Deniz; Sengul, GokhanIndustry requires low-cost, low-power consumption, and autonomous remote sensing systems for detecting and identifying magnetic materials. Magnetic anomaly detection is one of the methods that meet these requirements. This paper aims to detect and identify magnetic materials by the use of magnetic anomalies of the Earth's magnetic field created by some buried materials. A new measurement system that can determine the images of the upper surfaces of buried magnetic materials is developed. The system consists of a platform whose position is automatically controlled in x-axis and y-axis and a KMZ51 anisotropic magneto-resistive sensor assembly with 24 sensors mounted on the platform. A new identification system based on scale-invariant feature transform (SIFT)-binary robust invariant scalable keypoints (BRISKs) as keypoint and descriptor, respectively, is developed for identification by matching the similar images of magnetic anomalies. The results are compared by the conventional principal component analysis and neural net algorithms. On the six selected samples and the combinations of these samples, 100% correct classification rates were obtained.Article Citation - WoS: 19Citation - Scopus: 24Gender Detection Using 3d Anthropometric Measurements by Kinect(Polska Akad Nauk, Polish Acad Sciences, 2018) Camalan, Seda; Sengul, Gokhan; Misra, Sanjay; Maskeliunas, Rytis; Damasevicius, RobertasAutomatic gender detection is a process of determining the gender of a human according to the characteristic properties that represent the masculine and feminine attributes of a subject. Automatic gender detection is used in many areas such as customer behaviour analysis, robust security system construction, resource management, human-computer interaction, video games, mobile applications, neuro-marketing etc., in which manual gender detection may be not feasible. In this study, we have developed a fully automatic system that uses the 3D anthropometric measurements of human subjects for gender detection. A Kinect 3D camera was used to recognize the human posture, and body metrics are used as features for classification. To classify the gender, KNN, SVM classifiers and Neural Network were used with the parameters. A unique dataset gathered from 29 female and 31 male (a total of 60 people) participants was used in the experiment and the Leave One Out method was used as the cross-validation approach. The maximum accuracy achieved is 96.77% for SVM with an MLP kernel function.Article Citation - WoS: 11Citation - Scopus: 13Gesture-Based Interaction for Learning: Time To Make the Dream a Reality(Wiley, 2012) Ozcelik, Erol; Sengul, Gokhan[No Abstract Available]
