HivNet: Studying in depth the morphology of HIV-1 virion using Deep Learning

The capacity of transmission electron microscopy (TEM) to distinguish ultrastructure morphology at the nanometer scale makes it useful for a wide range of biomedical imaging applications. TEM has long been a vital tool in the virologist's toolbox because of its capacity to directly visualize virus particles. When used in HIV-1 research, TEM is essential for assessing the actions of inhibitors that obstruct the maturation and morphogenesis phases of the virus lifecycle. However, TEM micrograph fabrication and analysis both involve tedious manual effort. We have built an 8-layer convolutional neural network backbone capable of categorizing HIV-1 virions at various phases of maturity and morphogenesis via the devoted application of computer vision frameworks and machine learning techniques. On a wide range of micrographs made up of various experimental samples and magnifications, our results surpassed both typical CNN backbones and deep residual networks, obtaining 91.33 percent testing accuracy and 85.83 percent validation accuracy. We anticipate that this tool will be useful to a variety of studies.