Mixed-Method Quantization of Convolutional Neural Networks on an Adaptive FPGA Accelerator | IEEE Conference Publication | IEEE Xplore