COVID-19 detection using hybrid deep learning model in chest x-rays images

The novel-corona-virus is presently accountable for 547,782 deaths worldwide. It was first observed in China in late 2019 and, the increase in number of its affected cases seriously disturbed almost every nation in terms of its economical, structural, educational growth. Furthermore, with the advancement of data-analytics and machine learning towards enhanced diagnostic tools for the infection, the growth rate in the affected patients has reduced considerably, thereby making it critical for AI researchers and experts from medical radiology to put more efforts in this side. In this regard, we present a controlled study which provides analysis of various potential possibilities in terms of detection models/algorithms for COVID-19 detection from radiology-based images like chest x-rays. We provide a rigorous comparison between the VGG16, VGG19, Residual Network, Dark-Net as the foundational network with the Single Shot MultiBox Detector (SSD) for predictions. With some preprocessing techniques specific to the task like CLAHE, this study shows the potential of the methodology relative to the existing techniques. The highest of all precision and recall were achieved with DenseNet201 + SSD512 as 93.01 and 94.98 respectively.

Covid-19 detection from chest X-Ray images using advanced deep learning techniques

Like the Covid-19 pandemic, smallpox virus infection broke out in the last century, wherein 500 million deaths were reported along with enormous economic loss. But unlike smallpox, the Covid-19 recorded a low exponential infection rate and mortality rate due to advancement in medical aid and diagnostics. Data analytics, machine learning, and automation techniques can help in early diagnostics and supporting treatments of many reported patients. This paper proposes a robust and efficient methodology for the early detection of COVID-19 from Chest X-Ray scans utilizing enhanced deep learning techniques. Our study suggests that using the Prediction and Deconvolutional Modules in combination with the SSD architecture can improve the performance of the model trained at this task. We used a publicly open CXR image dataset and implemented the detection model with task-specific pre-processing and near 80:20 split. This achieved a competitive specificity of 0.9474 and a sensibility/accuracy of 0.9597, which shall help better decision-making for various aspects of identification and treat the infection. © 2021 Tech Science Press. All rights reserved.