Efficient Computation of Depthwise Separable Convolution in MoblieNet Deep Neural Network Models | IEEE Conference Publication | IEEE Xplore