Detection of fake face images using lightweight convolutional neural networks with stacking ensemble learning method.

Images and videos containing fake faces are the most common type of digital manipulation. Such content can lead to negative consequences by spreading false information. The use of machine learning algorithms to produce fake face images has made it challenging to distinguish between genuine and fake content. Face manipulations are categorized into four basic groups: entire face synthesis, face identity manipulation (deepfake), facial attribute manipulation and facial expression manipulation. The study utilized lightweight convolutional neural networks to detect fake face images generated by using entire face synthesis and generative adversarial networks. The dataset used in the training process includes 70,000 real images in the FFHQ dataset and 70,000 fake images produced with StyleGAN2 using the FFHQ dataset. 80% of the dataset was used for training and 20% for testing. Initially, the MobileNet, MobileNetV2, EfficientNetB0, and NASNetMobile convolutional neural networks were trained separately for the training process. In the training, the models were pre-trained on ImageNet and reused with transfer learning. As a result of the first trainings EfficientNetB0 algorithm reached the highest accuracy of 93.64%. The EfficientNetB0 algorithm was revised to increase its accuracy rate by adding two dense layers (256 neurons) with ReLU activation, two dropout layers, one flattening layer, one dense layer (128 neurons) with ReLU activation function, and a softmax activation function used for the classification dense layer with two nodes. As a result of this process accuracy rate of 95.48% was achieved with EfficientNetB0 algorithm. Finally, the model that achieved 95.48% accuracy was used to train MobileNet and MobileNetV2 models together using the stacking ensemble learning method, resulting in the highest accuracy rate of 96.44%.

A Novel Approach to Skin Lesion Segmentation: Multipath Fusion Model with Fusion Loss.

Segmentation of skin lesions plays a very important role in the early detection of skin cancer. However, indistinguishability due to various artifacts such as hair and contrast between normal skin and lesioned skin is an important challenge for specialist dermatologists. Computer-aided diagnostic systems using deep convolutional neural networks are gaining importance in order to cope with difficulties. This study focuses on deep learning-based fusion networks and fusion loss functions. For the automatic segmentation of skin lesions, U-Net (U-Net + ResNet 2D) with 2D residual blocks and 2D volumetric convolutional neural networks were fused for the first time in this study. Also, a new fusion loss function is proposed by combining Dice Loss (DL) and Focal Tversky Loss (FTL) to make the proposed fused model more robust. Of the 2594 image dataset, 20% is reserved for test data and 80% for training data. In test data training, a Jaccard score of 0.837 and a dice score of 0.918 were obtained. The proposed model was also scored on the ISIC 2018 Task 1 test images, whose ground truths were not shared. The proposed model performed well and achieved a Jaccard index of 0.800 and a dice score of 0.880 in the ISIC 2018 Task 1 test set. In addition, it has been observed that the new fused loss function obtained by fusing Focal Tversky Loss and Dice Loss functions in the proposed model increases the robustness of the model in the tests. The proposed new loss function fusion model has outstripped the cutting-edge approaches in the literature.

An Effective Method for Detecting and Classifying Diabetic Retinopathy Lesions Based on Deep Learning.

Diabetic retinopathy occurs as a result of the harmful effects of diabetes on the eyes. Diabetic retinopathy is also a disease that should be diagnosed early. If not treated early, vision loss may occur. It is estimated that one third of more than half a million diabetic patients will have diabetic retinopathy by the 22nd century. Many effective methods have been proposed for disease detection with deep learning. In this study, unlike other studies, a deep learning-based method has been proposed in which diabetic retinopathy lesions are detected automatically and independently of datasets, and the detected lesions are classified. In the first stage of the proposed method, a data pool is created by collecting diabetic retinopathy data from different datasets. With Faster RCNN, lesions are detected, and the region of interests are marked. The images obtained in the second stage are classified using the transfer learning and attention mechanism. The method tested in Kaggle and MESSIDOR datasets reached 99.1% and 100% ACC and 99.9% and 100% AUC, respectively. When the obtained results are compared with other results in the literature, it is seen that more successful results are obtained.

The adaptive ARMA analysis of EMG signals.

In this study, Adaptive auto regressive-moving average (A-ARMA) analysis of EMG signals recorded on the ulnar nerve region of the right hand in resting position was performed. A-ARMA method, especially in the calculation of the spectrums of stationary signals, is used for frequency analysis of signals, which give frequency response as sharp peaks and valleys. In this study, as the result of A-ARMA method analysis of EMG signals frequency-time domain, frequency spectrum curves (histogram curves) were obtained. As the images belonging to these histograms were evaluated, fibrillation potential widths of the muscle fibers of the ulnar nerve region of the people (material of the study) were examined. According to the degeneration degrees of the motor nerves, 22 people had myopathy, 43 had neuropathy, and 28 were normal.

Prediction of minor head injured patients using logistic regression and MLP neural network.

In this study it is aimed to assess the posttraumatic cerebral hemodynamia in minor head injured patients. Eighty patients with minor head injury (Group 1) evaluated in the early 8 h of posttraumatic period between July 2003 and February 2004. The control group (Group 2) has composed of 32 healthy people. Bilateral blood flow velocities of middle cerebral arteries (MCA) had measured using transtemporal technique while internal carotid arteries were evaluated by submandibular examination. Two different mathematical models such as the traditional statistical method on the basis of logistic regression and a multi-layer perceptron (MLP) neural network are used to classify the age, sex, velocitiy parameters of MCA, mean velocity of extracranial ICAs and V(MCA)/ V(ICA) ratios. The neural network was trained, cross-validated and tested with subject's transcranial Doppler signals. As a result of these classifications, we found the success rate of logistic regression, the success rate of MLP neural network is 88.2 and 89.1%, respectively. The classification results show that MLP neural network is offering the best results in the case of diagnosis.

Examination of the effects of degeneration on vertebral artery by using neural network in cases with cervical spondylosis.

The scope of this study is to diagnose vertebral arterial inefficiency by using Doppler measurements from both right and left vertebral arterials. Total of 96 patients' Doppler measurements, consisting of 42 of healthy, 30 of spondylosis, and 24 of clinically proven vertebrobasillary insufficiency (VBI), were examined. Patients' age and sex information as well as RPSN, RPSVN, LPSN, LPSVN, and TOTALVOL medical parameters obtained from vertebral arterials were classified by neural networks, and the performance of said classification reached up to 93.75% in healthy, 83.33% in spondylosis, and 97.22% in VBI cases. The area under ROC curve, which is a direct indication of repeating success ratio, is calculated as 92.3%, and the correlation coefficient of the classification groups is 0.9234. It is also demonstrated that those medical parameters of age and systolic velocity, which were applied into the neural networks, were more effective in developing vertebral deficiency.

The examination of the effects of obesity on a number of arteries and body mass index by using expert systems.

In this study, the areas affected from obesity were examined by classifying divergent arteries and body mass index (BMI) of 30 healthy persons and 52 obese persons by using expert systems, and the classifying performances of NEFCLASS and CANFIS, which are expert systems were compared. As a result of this comparison, it is observed that the classifying performance of NEFCLASS is better than that of CANFIS, and the causes of this are examined. Furthermore, it is observed that after these classifications, obesity affects the BMI rather than divergent arteries.

Application of FFT analyzed cardiac Doppler signals to fuzzy algorithm.

Doppler signals, recorded from the output of tricuspid, mitral, and aorta valves of 60 patients, were transferred to a personal computer via 16-bit sound card. The fast Fourier transform (FFT) method was applied to the recorded signal from each patient. Since FFT method inherently cannot offer a good spectral resolution at highly turbulent blood flows, it sometimes leads to wrong interpretation of cardiac Doppler signals. In order to avoid this problem, firstly six known diseased heart signals such as hypertension, mitral stenosis, mitral failure, tricuspid stenosis, aorta stenosis, aorta insufficiency were introduced to fuzzy algorithm. Then, the unknown heart diseases from 15 patients were applied to the same fuzzy algorithm in order to detect the kinds of diseases. It is observed that the fuzzy algorithm gives true results for detecting the kind of diseases.