Online ISSN: 2515-8260

Keywords : AlexNet


Automatic Classification of the Severity of COVID-19 Patients Based on CT Scans and X-rays Using Deep Learning

Sara Bhatti; Dr. Asif Aziz; Dr. Naila Nadeem; Irfan Usmani; Prof. Dr. Muhammad Aamir; Dr. Irum Khan

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 10, Pages 1436-1455

The 2019 novel coronavirus (COVID-19), which originated from China, has been declared a pandemic by the World Health Organization (WHO) as it has surpassed over eighty three million cases worldwide, with nearly two million deaths. The unexpected exponential increase in positive cases and the limited number of ventilators, personal safety equipment and COVID-19 test kits, especially in Low to Middle Income Countries (LMIC), had put undue pressure on medical staff, first responders as well as the overall health care systems. The Real-Time Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR) test is the decisive test for the diagnosis of COVID-19, but a significant percentage of positive tests return a false negative result. For patients in LMICs, the availability and affordability of routine Computerized Tomography (CT) scanning and chest X-rays is better compared to an RT-PCR test, especially in rural areas. Chest X-rays and CT scans can aid in the prognosis and management of COVID-19 positive patients, but are not recommended for diagnostic purposes. Using Deep Convolutional Neural Networks (CNN), three network based pre-trained models (AlexNet, GoogleNet and Resnet50) were used for the automatic classification of positive COVID-19 chest X-Rays and CT scans based on their severity into three classes- normal, mild/moderate, severe. This classification can aid health care workers in performing expeditious analysis of large numbers of thoracic CT scans and chest X-rays of COVID-19 positive patients, and aid in their prognosis and management. The images were obtained from public repositories, and were verified and classified by trained and highly experienced radiologist from Agha Khan University Hospital prior to simulations. The images were augmented and trained, and ResNet50 was concluded to achieve the highest accuracy. This research can be used for other lung infections, and can aid the authorities in the preparations of future pandemics.

Brain Tumor And Intracranial Haemorrhage Feature Extraction And Classification Using Conventional And Deep Learning Methods

R. Aruna Kirithika; S. Sathiya; M. Balasubramanian; P. Sivaraj

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 7, Pages 237-258

Presently, brain tumor (BT) and Intracranial hemorrhage (ICH) detection and classification processes become essential to save human lives. Automated diagnosis model using deep learning (DL) models finds useful to attain improved diagnostic outcome. This paper presents an ensemble of handcrafted and deep features for BT and ICH diagnosis. The proposed model comprises of three important processes, such as preprocessing, feature extraction and classification. The preprocessing of the input image takes place in three ways namely skull stripping, bilateral filtering (BF) and contrast limited adaptive histogram equalization (CLAHE) based contrast enhancement. In addition, scale invariant feature transform (SIFT) and AlexNet models are used for feature extraction process. In order to classify the existence of BT and ICH, two classification models is carried out such as gaussian naïve bayes (GNB) and random forest (RF).For validating the effective diagnostic performance of the proposed model, a set of simulations were carried out to determine the different class labels. The simulation outcome indicated the effective performance with the maximum sensitivity of 92.41%, specificity of 100%, and accuracy of 94.26%.

Helmet, Violation, Detection Using Deep Learning

Sherin Eliyas; K. Swaathi; Dr.P. Ranjana; A. Harshavardhan

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 2, Pages 5173-5178

Road incidents are among the significant reasons, for the human passing. The majority of the passings in mishaps are because of harm to the top of the bike riders. Among the various sorts of street mishaps, bike mishaps are normal and cause extreme wounds. To reduce the involved risk for the motorcycle riders it is exceptionally fascinating to utilize helmet. The helmet is the motorcyclist's primary security. Many countries require the utilization of caps by motorcyclists, however numerous individuals neglect to comply with the law for different reasons. We present the advancement of a framework utilizing profound convolutional neural networks, (CNNs) for discovering bikers who are disregarding cap rules. The system involves motorcycle, detection, helmet, vs. no-helmet, classification, and method counting. Faster R-CNN with ResNet 50 network, model is implementing for motorcycle detector process. CNN classification model proposes for classify the helmet vs. no-helmet. Finally making alarm sound to alert the officer too preventing motorcycle accident. We assess the framework as far as accuracy and speed.

Automated Identification of Glaucoma from Fundus Images using Deep learning Techniques

Ajitha S; Dr. M V Judy; Dr. Meera N; Dr. Rohith N

European Journal of Molecular & Clinical Medicine, 2020, Volume 7, Issue 2, Pages 5449-5458

Glaucoma has arisen as the one of the main sources of visual impairment. A typical technique for diagnosing glaucoma is through assessment optic nerve head by an experienced ophthalmologist. This methodology is arduous and burns-through a lot of time. Despite the fact that the analysis of this infection has not yet been discovered, the period of primary identification can preserve from the glaucoma. Subsequently, customary glaucoma screening is basic and suggested. The issue can be settled by applying machine learning techniques for glaucoma detection. We present an automated glaucoma screening framework using a pre-trained Alexnet model with SVM classifier to enhance the classification accuracy . In this study, we used three publicly available dataset as HRF, Origa and Drishti_GS1 dataset. The proposed model achieved the image classification accuracy of 91.21%. This study showed that using pre-trained CNN with SVM for glaucoma detection showed greater accuracy in automatic image classification than just CNN or SVM.