Efficient HLS Implementation for Convolutional Neural Networks Accelerator on an SoC | IEEE Conference Publication | IEEE Xplore