Scopus İndeksli Yayınlar Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/11727/4809
Browse
3 results
Search Results
Item 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-8820-2019Femoral 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.Item 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-8820-2019Objectives: 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.Item Integrating features for accelerometer-based activity recognition(2016) Erdas, C.Berke; Atasoy, Isil; Acici, Koray; Ogul, Hasan; 0000-0003-3467-9923Activity recognition is the problem of predicting the current action of a person through the motion sensors worn on the body. The problem is usually approached as a supervised classification task where a discriminative model is learned from known samples and a new query is assigned to a known activity label using learned model. The challenging issue here is how to feed this classifier with a fixed number of features where the real input is a raw signal of varying length. In this study, we consider three possible feature sets, namely time-domain, frequency domain and wavelet-domain statistics, and their combinations to represent motion signal obtained from accelerometer reads worn in chest through a mobile phone. In addition to a systematic comparison of these feature sets, we also provide a comprehensive evaluation of some preprocessing steps such as filtering and feature selection. The results determine that feeding a random forest classifier with an ensemble selection of most relevant time-domain and frequency-domain features extracted from raw data can provide the highest accuracy in a real dataset. (C) 2016 The Authors. Published by Elsevier B.V.