| Title |
Implementation of SNN/CNN Accumulator H/W using LIF/IF Model |
| Authors |
홍윤표(Yun-Pyo Hong) ; 김희탁(Hee-Tak Kim) ; 전석훈(Seok-Hun Jeon) ; 황태호(Tae-Ho Hwang) |
| DOI |
https://doi.org/10.5573/ieie.2022.59.1.112 |
| Abstract |
This paper explains h/w implementation of Integrated and Fired (IF) & Leaky IF (LIF) Neuron Model for SNN, and propses an accumulator for both a spiking neural network (SNN) and a convolution neural network (CNN). It is implemented with minimal hardwarer size increase using the similarity between activation function of CNN and neuron model of SNN. In addition, we compare the image classification accuracy for IF and LIF model using CIFAR-10 Dataset. As a result, we get 89.8% of image classification accuracy using LIF model, and make the accumulator for both with 31% increase of h/w size compared to the accumulator only for the CNN. |