A Block-Floating-Point Arithmetic Based FPGA Accelerator for Convolutional Neural Networks | IEEE Conference Publication | IEEE Xplore