Block Convolution: Toward Memory-Efficient Inference of Large-Scale CNNs on FPGA | IEEE Journals & Magazine | IEEE Xplore