Applying Convolutional Neural Networks to Identify Parasitized Malaria Cells

“We develop an algorithm that can detect malaria from images of segmented cells from the thin blood smear slide images with 96% accuracy. Our algorithm, SimpNet-7, is a 7-layer convolutional neural network trained on the NIH Malaria dataset, containing 27,588 images of parasitized and uninfected cells. We find that SimpNet-7 achieves an F1 score of 0.955, a precision of 0.946, and a recall of 0.974. We then propose the application of this algorithm in hospitals in areas where malaria is prominent but medical resources are sparse, such as African countries.”

A Comparative Study of Deep Learning Models to Identify Tuberculosis and Pneumonia from Chest X-Rays

“These CNNs may aid radiologists in delivering diagnoses more accurately and quickly. Testing that was completed evaluated the accuracy of the InceptionResNetV2, ResNet50, ResNet152, DenseNet121, InceptionV3, and AlexNet models in creating a multi-class classification model to differentiate between CXRs showing TB, bacterial pneumonia, and normal lungs. The generated results showcase how differing architectures and abilities of each individual model uniquely influenced their decision-making within the three classes of CXRs…”

A Fast Machine Learning Model for ECG-Based Heartbeat Classification and Arrhythmia Detection

“We present a fully automatic and fast ECG arrhythmia classifier based on a simple brain-inspired machine learning approach known as Echo State Networks. Our classifier has a low-demanding feature processing that only requires a single ECG lead. Its training and validation follows an inter-patient procedure. Our approach is compatible with an online classification that aligns well with recent advances in health-monitoring wireless devices and wearables…”

Effect of the Laplace Operator on the Performance of CoronaNet for COVID-19 Detection

“Coronavirus disease (COVID-19) is the cause of a global pandemic that is affecting millions of people around the world. Inadequate testing resources have resulted in several people going undiagnosed and consequently untreated. However, using computerized tomography (CT) scans for diagnosis is an alternative to bypass this limitation. In this paper, we describe CoronaNet, a deep convolutional neural network that can recognize if a patient has COVID-19 from images of CT scans with 91% accuracy. We hope this algorithm can be incorporated into hospitals around the world to assist in coronavirus mitigation efforts. Additionally, we demonstrate the effectiveness of the Laplace Operator in enhancing the performance of CoronaNet. “