TY - GEN
T1 - Regularized Pooling
AU - Otsuzuki, Takato
AU - Hayashi, Hideaki
AU - Zheng, Yuchen
AU - Uchida, Seiichi
N1 - Funding Information:
Acknowledgments. This work was supported by JSPS KAKENHI Grant JP17H06100 and JST ACT-I Grant Number JPMJPR18UO.
Publisher Copyright:
© 2020, Springer Nature Switzerland AG.
PY - 2020
Y1 - 2020
N2 - In convolutional neural networks (CNNs), pooling operations play important roles such as dimensionality reduction and deformation compensation. In general, max pooling, which is the most widely used operation for local pooling, is performed independently for each kernel. However, the deformation may be spatially smooth over the neighboring kernels. This means that max pooling is too flexible to compensate for actual deformations. In other words, its excessive flexibility risks canceling the essential spatial differences between classes. In this paper, we propose regularized pooling, which enables the value selection direction in the pooling operation to be spatially smooth across adjacent kernels so as to compensate only for actual deformations. The results of experiments on handwritten character images and texture images showed that regularized pooling not only improves recognition accuracy but also accelerates the convergence of learning compared with conventional pooling operations.
AB - In convolutional neural networks (CNNs), pooling operations play important roles such as dimensionality reduction and deformation compensation. In general, max pooling, which is the most widely used operation for local pooling, is performed independently for each kernel. However, the deformation may be spatially smooth over the neighboring kernels. This means that max pooling is too flexible to compensate for actual deformations. In other words, its excessive flexibility risks canceling the essential spatial differences between classes. In this paper, we propose regularized pooling, which enables the value selection direction in the pooling operation to be spatially smooth across adjacent kernels so as to compensate only for actual deformations. The results of experiments on handwritten character images and texture images showed that regularized pooling not only improves recognition accuracy but also accelerates the convergence of learning compared with conventional pooling operations.
UR - http://www.scopus.com/inward/record.url?scp=85094148287&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85094148287&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-61616-8_20
DO - 10.1007/978-3-030-61616-8_20
M3 - Conference contribution
AN - SCOPUS:85094148287
SN - 9783030616151
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 241
EP - 254
BT - Artificial Neural Networks and Machine Learning – ICANN 2020 - 29th International Conference on Artificial Neural Networks, Proceedings
A2 - Farkaš, Igor
A2 - Masulli, Paolo
A2 - Wermter, Stefan
PB - Springer Science and Business Media Deutschland GmbH
T2 - 29th International Conference on Artificial Neural Networks, ICANN 2020
Y2 - 15 September 2020 through 18 September 2020
ER -