Fish remains popular among the body’s most essential nutrients, as it contains protein and polyunsaturated fatty acids. It is extremely important to choose the fish consumption according to the season and the freshness of the fish to be purchased. It is very difficult to distinguish between non-fresh fish and fresh fish mixed in the fish stalls. In addition to traditional methods used to determine meat freshness, significant success has been achieved in studies on fresh fish detection with artificial intelligence techniques. In this study, two different types of fish (anchovy and horse mackerel) used to determine fish freshness with convolutional neural networks, one of the artificial intelligence techniques. The images of fresh fish were taken, images of non-fresh fish were taken and two new datasets (Dataset1: Anchovy, Dataset2: Horse mackerel) were created. A novel hybrid model structure has been proposed to determine fish freshness using fish eye and gill regions on these two datasets. In the proposed model, Yolo-v5 and Inception-ResNet-v2 and Xception model structures are used through transfer learning. Whether the fish is fresh in both of the Yolo-v5 + Inception-ResNet-v2 (Dataset1: 97.67%, Dataset2: 96.0%) and Yolo-v5 + Xception (Dataset1: 88.00%, Dataset2: 94.67%) hybrid models created using these model structures has been successfully detected. Thanks to the model we have proposed, it will make an important contribution to the studies that will be conducted in the freshness studies of fish using different storage days and the estimation of fish size.

No abstract

Recently, coronavirus disease (Covid-19) has become a serious public health threat, spreading worldwide in a very short time and threatening the lives of millions. With the increasing number of cases and mutations, medical resources are being drained day by day due to the rapid transmission of the disease, and the health systems of many countries are negatively affected. For this reason, it is very important to use available resources appropriately and timely for the detection and treatment of the disease. In this study, VGG16 and ResNet50 deep learning models were used to quickly evaluate x-ray images and to make the pre-diagnosis of Covid-19, and an alternative model (IsVoNet) was proposed. As a result of the training of the models, success accuracy of 99.92% in the VGG16 model, 99.65% in the ResNet50 model and 99.76% in the proposed model were obtained. According to the results, it was observed that the models classified Covid-19 and normal lung x-ray images with high accuracy and the proposed model showed a high success rate at lower time complexity than other models.

Skin cancer is one of the most common type of cancer in humans. This type of cancer is produced by skin cells called melanocytes and occurs as a result of division and multiplication of the mentioned cells. The most important symptom of skin cancer is the formation of spots on the skin or the observation of changes in the shape, color, or size of the existing spot. It is necessary to consult a specialist to distinguish the difference between a normal spot and skin cancer. Expert physicians examine and follow up the spots on the skin using skin surface microscopy, called dermatoscopy, or take a sample from the suspicious area and request it to be examined in laboratory environment. This situation increases the cost of the procedure for the diagnosis of skin cancer and also causes it to be treated at a later stage. Therefore, there is a need for a metod that can detect skin cancer early. Thanks to machine learning, become popular in recent years, many diseases can be diagnosed with software that helps expert physicians. In this study, VGGNet model structures (VGG-11, VGG-13, VGG-16, VGG-19) that quickly classify skin cancer and become a traditional convolutional neural network architecture using deep learning method, a subfield of machine learning, were used. It has been observed that the VGG-11 architecture, which is one of the VGGNet model structures, detects skin cancer with superior success accuracy (83%) compared to other model structures.

Emotion recognition from facial expression is a current research topic that can be applied in the many fields of computer vision, such as human-computer interaction, emotional computing, etc. In this study, an application for emotion recognition through deep learning was made using KDEF and PICS datasets. A new model was established using the convolutional neural network (CNN), an artificial intelligence approach that involves artificial neural networks, which is one of the deep learning techniques for attribute inference. Large datasets are needed for the high performance of deep learning. There are 4900 images in the KDEF dataset and 322 images in the PICS dataset. For this purpose, primarily due to the small number of images in the PICS dataset, image iteration was applied with the data augmentation method, and the PICS dataset was increased to 4830 images. Then, the new model developed by conducting separate training on these two different datasets was tested. Seven different classes of emotion (afraid, angry, disgusted, happy, neutral, sad, surprised) were covered in each dataset in the study conducted with a new model developed based on VGGNet which is one of the CNN models. With the developed model, a high success rate was achieved by obtaining 97.44% accuracy values in the validation set of the KDEF and 98.24% accuracy values in the validation set of the PICS dataset.

Printed Circuit Boards (PCBs) are electronic boards that hold electronic components together and provide the electrical connection between these components. Printed circuit boards offer many advantages over traditional wired circuits, such as durability, less heat, minimal wiring, and ease of assembly. Correct design and production of printed circuit boards significantly affect the quality and efficiency of printed circuit boards. In this study, a defect detection system based on machine learning and deep learning algorithms is proposed to help produce printed circuit boards accurately and minimize the error rate. In the proposed system, missing hole, mouse bite, open circuit, short, spur, and spurious copper defects on the printed circuit have been determined. According to the results obtained, According to the results obtained, success accuracies of 74.62% were obtained with YOLO-v4, 47.83% with HOG+SVM, and 39.86% with HOG+KNN. It has been seen that the algorithms discussed in the study are applicable in the detection of defects in printed circuit boards.

No abstract

Today, tick-borne diseases have become widespread and pose a significant threat to public health. There are various types of diseases transmitted from ticks to humans. The main of these diseases; Crimean-Congo hemorrhagic fever, Lyme disease, Mediterranean spotted fever and Tularemia can be listed. As with other types of diseases, early diagnosis is important in tick-borne diseases. Therefore, it is necessary to identify ticks quickly and accurately in order to reduce the possible risks of disease in cases of errors bitten by ticks. In this study, Yolo-v3-based deep learning algorithm, which is a subfield of machine learning, was used primarily to detect ticks. For the training and testing of this algorithm, a new data set was created by downloading 1500 different tick images from the internet. Algorithm was trained and tested using this data set. In order to determine that the success accuracy of the Yolo-v3-based deep learning algorithm is superior and to demonstrate its availability in real life, various performance tests were performed and an estimate was made as to whether there were ticks in an image. As a result of the study, in order to reduce the disease risk of patients bitten by ticks and to intervene in a timely manner, tick detection was made effectively by taking only one tick picture.