Automated detection of Covid-19 disease using deep fused features from chest radiography images.

The health systems of many countries are desperate in the face of Covid-19, which has become a pandemic worldwide and caused the death of hundreds of thousands of people. In order to keep Covid-19, which has a very high propagation rate, under control, it is necessary to develop faster, low-cost and highly accurate methods, rather than a costly Polymerase Chain Reaction test that can yield results in a few hours. In this study, a deep learning-based approach that can detect Covid-19 quickly and with high accuracy on X-ray images, which are common in every hospital and can be obtained at low cost, was proposed. Deep features were extracted from X-Ray images in RGB, CIE Lab and RGB CIE color spaces using DenseNet121 and EfficientNet B0 pre-trained deep learning architectures and then obtained features were fed into a two-stage classifier approach. Each of the classifiers in the proposed approach performed binary classification. In the first stage, healthy and infected samples were separated, and in the second stage, infected samples were detected as Covid-19 or pneumonia. In the experiments, Bi-LSTM network and well-known ensemble approaches such as Gradient Boosting, Random Forest and Extreme Gradient Boosting were used as the classifier model and it was seen that the Bi-LSTM network had a superior performance than other classifiers with 92.489% accuracy.

Learn from one data set to classify all - A multi-target domain adaptation approach for white blood cell classification.

BACKGROUND AND OBJECTIVE: Traditional machine learning methods assume that both training and test data come from the same distribution. In this way, it becomes possible to achieve high successes when modelling on the same domain. Unfortunately, in real-world problems, direct transfer between domains is adversely affected due to differences in the data collection process and the internal dynamics of the data. In order to cope with such drawbacks, researchers use a method called "domain adaptation", which enables the successful transfer of information learned in one domain to other domains. In this study, a model that can be used in the classification of white blood cells (WBC) and is not affected by domain differences was proposed. METHODS: Only one data set was used as source domain, and an adaptation process was created that made possible the learned knowledge to be used effectively in other domains (multi-target domain adaptation). While constructing the model, we employed data augmentation, data generation and fine-tuning processes, respectively. RESULTS: The proposed model has been able to extract "domain-invariant" features and achieved high success rates in the tests performed on nine different data sets. Multi-target domain adaptation accuracy was measured as %98.09. CONCLUSIONS: At the end of the study, it has been observed that the proposed model ignores the domain differences and it can adapt in a successful way to target domains. In this way, it becomes possible to classify unlabeled samples rapidly by using only a few number of labeled ones.

Classification of white blood cells using capsule networks.

BACKGROUND: While the number and structural features of white blood cells (WBC) can provide important information about the health status of human beings, the ratio of sub-types of these cells and the deformations that can be observed serve as a good indicator in the diagnosis process of some diseases. Hence, correct identification and classification of the WBC types is of great importance. In addition, the fact that the diagnostic process that is carried out manually is slow, and the success is directly proportional to the expert's skills makes this problem an excellent field of application for computer-aided diagnostic systems. Unfortunately, both the ethical reasons and the cost of image acquisition process is one of the biggest obstacles to the fact that researchers working with medical images are able to collect enough data to produce a stable model. For that reasons, researchers who want to perform a successful analysis with small data sets using classical machine learning methods need to undergo their data a long and error-prone pre-process, while those using deep learning methods need to increase the data size using augmentation techniques. As a result, there is a need for a model that does not need pre-processing and can perform a successful classification in small data sets. METHODS: WBCs were classified under five categories using a small data set via capsule networks, a new deep learning method. We improved the model using many techniques and compared the results with the most known deep learning methods. RESULTS: Both the above-mentioned problems were overcame and higher success rates were obtained compared to other deep learning models. While, convolutional neural networks (CNN) and transfer learning (TL) models suffered from over-fitting, capsule networks learned well training data and achieved a high accuracy on test data (96.86%). CONCLUSION: In this study, we briefly discussed the abilities of capsule networks in a case study. We showed that capsule networks are a quite successful alternative for deep learning and medical data analysis when the sample size is limited.

Classification of DNA damages on segmented comet assay images using convolutional neural network.

BACKGROUND AND OBJECTIVE: Identification and quantification of DNA damage is a very significant subject in biomedical research area which still needs more robust and effective methods. One of the cheapest, easy to use and most successful method for DNA damage analyses is comet assay. In this study, performance of Convolutional Neural Network was examined on quantification of DNA damage using comet assay images and was compared to other methods in the literature. METHODS: 796 single comet grayscale images with 170 x 170 resolution labeled by an expert and classified into 4 classes each having approximately 200 samples as G0 (healthy), G1 (poorly defective), G2 (defective) and G3 (very defective) were utilized. 120 samples were used as test dataset and the rest were used in data augmentation process to achieve better performance with training of Convolutional Neural Network. The augmented data having a total of 9995 images belonging to four classes were used as network training data set. RESULTS: The proposed model, which was not dependent to pre-processing parameters of image processing for DNA damage classification, was able to classify comet images into 4 classes with an overall accuracy rate of 96.1%. CONCLUSIONS: This paper primarily focuses on features and usage of Convolutional Neural Network as a novel method to classify comet objects on segmented comet assay images.