Hybrid Approach for Efficient Quantization of Weights in Convolutional Neural Networks | IEEE Conference Publication | IEEE Xplore