To describe a model using classic machine learning techniques for creating machine learning systems, a person who specializes in this technique needs to extract feature vectors. This period also breaks into expert time. Also, these methods could not process raw data without preprocessing and expert assistance. Deep learning has made great progress in solving problems at this point, and machine learning research has continued for many years. Unlike traditional machine learning and image processing techniques, deep networks enable learning processes using raw data. In this study, a deep learning approach for the classification and diagnosis of malaria is developed. For this purpose, Residual Attention Network (RAN) a deep learning Convolutional Neural Network (CNN) technique was used with previously classified datasets. The goal is to design computer-aided software for classifying blood cell images (blood samples) as "parasitized" or "uninfected". In the program, a decision support system was implemented by a deep learning approach. As a result, the RAN model achieved the best ability to produce better results in processing and classification images compared to other algorithm types. RAN model's training simulation results showed a 95.79% classification accuracy rate. Using the Support Vector Machine (SVM) obtained only 83.30% classification accuracy rate. Besides, it is evaluated that for the classification of blood cell images and diagnosis of malaria using deep learning methods can be used successfully. In addition, deep learning methods have the advantage of automatically learning features from input data and require minimal input by specialists in automated malaria diagnosis.

The identification of leucocyte, also named white blood cells, types in histological blood tissue images is significant because it enables an opportunity for the diagnosis of various hematological diseases. In this study, for the diagnosis of lymphoma cancer, a hematologic disorder, we presented automatic detection and classification model using a deep learning approach. Faster R-CNN, which is a kind of region-based Convolutional Neural Network (CNN) model, achieves satisfactory performance on object detection and classification problems. To dispose of the feature extraction process in image-based applications, we offer a ResNet50 modified Faster R-CNN model for the detection and classification of leucocyte types which are lymphocyte, monocyte, basophil, eosinophil, and neutrophil in histological blood tissue images. In parallel with this purpose, a novel Faster R-CNN object detection model was designed by modifying ResNet50 model and the locations of leucocytes in the image were determined and classified. The efficiency of the proposed model was tested on a novel histological dataset including blood tissue images. The number of lymphocytes in the blood tissue is used as an evaluation criterion in the diagnosis of lymphoma cancer. Therefore, this study sets an example for clinical studies. According to the proposed model, firstly, the blood tissue images are normalized, and the implicit features are extracted by using the trainable convolution kernel. Then, for the reduction of the extracted implicit features, the maximum pooling is applied. After that, Region Proposal Networks (RPNs) are used to generate high-quality region proposals, which are used by Faster R-CNN for detection. Finally, the softmax classifier and regression layer are carried out to categorize the leucocyte types and estimate the boundary boxes of the test samples, respectively. Experimental results show the successful performance and the generalization capability of novel Faster R-CNN for the detection and classification of leucocyte types. This model demonstrates the potential to be deployed as a diagnostic tool for clinical studies because the method has been tested on a real-world histological data set.

No abstract

Abstract:In educational activities, examinations are sometimes carried out in the form of multiple-choice tests or sometimes as open-ended long texts. When multiple-choice tests are performed, evaluating process is carried out either manual or computer-assisted. Exam questions prepared in the form of multiple choice tests are not suitable for every course. It may be necessary to use open-ended questionnaires in order for pupils to accurately measure their achievement in relation to the course. It can take a long time to evaluate examinations made with such questions. However, this process can create problems in terms of objective evaluation. Data mining, defined as the extraction of useful information from large quantities of data, can be used to process all kinds of data. The data mining method used in the processing of textual data is called text mining. In text processing studies, data is subject to preprocessing in order to obtain a high quality data set. The most important stage of preprocessing is stemming. In this study, stemming process is implemented to questions and correct answers taken from students. The results obtained in 2 different samples and 4 sentences are 71%, 69%, 86% and 78% correct. In order to be able to distinguish what the textual data written in the natural language really is, it is necessary to use the states of the words which are made up of construction and free from the suffixes. Therefore, in the pre-processing phase, stemming process is applied to the textual data in accordance with the grammar rules of the language they are written on, and stems of every word are found. Text processing is used in many areas of the natural language. Computer-aided solutions will be inevitable so that problems can be eliminated and open-ended questions can be quickly assessed. Despite the desirability of a computer aided solution for this measurement technique, studies of this solution are not included in the literature very much.