TY - CHAP
T1 - A Method for Reducing Number of Parameters of Octave Convolution in Convolutional Neural Networks
AU - Gotoh, Yusuke
AU - Inoue, Yu
N1 - Funding Information:
Acknowledgement. This work was supported by JSPS KAKENHI Grant Number 18K11265 and 21H03429, and JGC-S Scholarship Foundation.
Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - Computational machine learning has attracted a great deal of attention for its ability to analyze large-scale data. In particular, convolutional neural networks (CNNs) have been proposed in the fields of image recognition and object detection in efforts to develop models with improved accuracy as well as more lightweight models that require a smaller number of parameters and a lower computational cost. Octave Convolution (OctConv) is a method used to reduce the memory and computational cost of a model while also improving its accuracy by replacing the conventional convolutional layer with an OctConv layer. However, the number of parameters used in OctConv is almost the same as that in the case of conventional convolutional processing. In this paper, we propose the Pointwise Octave Convolution (Pointwise OctConv) method, which combines the Pointwise Convolution (Pointwise Conv) method with OctConv to reduce the number of parameters used in OctConv and thus create a lighter model. In the proposed method, the number of parameters is reduced by performing Pointwise Conv before and after the convolution process for each path in the OctConv layer. In an evaluation using ResNet-56, the proposed method reduces the number of parameters by about 63.8 % with a loss of classification accuracy of 3.04 % when α= 0.75.
AB - Computational machine learning has attracted a great deal of attention for its ability to analyze large-scale data. In particular, convolutional neural networks (CNNs) have been proposed in the fields of image recognition and object detection in efforts to develop models with improved accuracy as well as more lightweight models that require a smaller number of parameters and a lower computational cost. Octave Convolution (OctConv) is a method used to reduce the memory and computational cost of a model while also improving its accuracy by replacing the conventional convolutional layer with an OctConv layer. However, the number of parameters used in OctConv is almost the same as that in the case of conventional convolutional processing. In this paper, we propose the Pointwise Octave Convolution (Pointwise OctConv) method, which combines the Pointwise Convolution (Pointwise Conv) method with OctConv to reduce the number of parameters used in OctConv and thus create a lighter model. In the proposed method, the number of parameters is reduced by performing Pointwise Conv before and after the convolution process for each path in the OctConv layer. In an evaluation using ResNet-56, the proposed method reduces the number of parameters by about 63.8 % with a loss of classification accuracy of 3.04 % when α= 0.75.
UR - http://www.scopus.com/inward/record.url?scp=85124118888&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85124118888&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-95903-6_23
DO - 10.1007/978-3-030-95903-6_23
M3 - Chapter
AN - SCOPUS:85124118888
T3 - Lecture Notes on Data Engineering and Communications Technologies
SP - 212
EP - 222
BT - Lecture Notes on Data Engineering and Communications Technologies
PB - Springer Science and Business Media Deutschland GmbH
ER -