Accelerating Deep Neural Network Computation on a Low Power Reconfigurable Architecture | IEEE Conference Publication | IEEE Xplore