IEEE Transactions on Computers
Efficient SIMD implementation for accelerating convolutional neural network