VLSI implementation of a neural network classifier based on the saturating linear activation function | IEEE Conference Publication | IEEE Xplore