YodaNN: An Ultra-Low Power Convolutional Neural Network Accelerator Based on Binary Weights.

Renzo Andri Lukas Cavigelli Davide Rossi Luca Benini

Published in: ISVLSI (2016)

Keyphrases

convolutional neural network
ultra low power
face detection
low power
weighted sum
linear combination
weighted average
non binary
relative importance
data sets
weighting scheme
neural network
parallel implementation
database
hamming distance
multi class
training data
weight vector
image processing
binary data