XSepConv: extremely separated convolution for efficient deep networks with large kernels

Jiarong Chen; Zongqing Lu; Qingmin Liao

doi:10.1117/12.2601043

30 June 2021 XSepConv: extremely separated convolution for efficient deep networks with large kernels

Jiarong Chen, Zongqing Lu, Qingmin Liao

Proceedings Volume 11878, Thirteenth International Conference on Digital Image Processing (ICDIP 2021); 118780K (2021) https://doi.org/10.1117/12.2601043
Event: Thirteenth International Conference on Digital Image Processing, 2021, Singapore, Singapore

Abstract

Due to the increasing demand for deploying CNNs on resource-constrained platforms, efficient neural networks are becoming more and more popular, in which depthwise convolution plays an indispensable role. Recently, larger kernel sizes (≥5) have been applied to depthwise convolution, but with significantly increased computational cost and parameter size. In this paper, we propose a novel extremely separated convolutional block (XSepConv), which fuses spatially separable convolutions into depthwise convolution to substantially reduce both the computational cost and the parameter size induced by large kernels. We also propose an extra 2×2 depthwise convolution coupled with a new and improved symmetric padding strategy to compensate for the side effect brought by spatially separable convolutions. XSepConv is a more efficient alternative to vanilla depthwise convolution with large kernel sizes; moreover, extensive experiments on popular image classification and object detection benchmarks demonstrate that XSepConv can strike a better trade-off between accuracy and efficiency, and the improvement is more significant for larger kernels.

Citation Download Citation

Jiarong Chen, Zongqing Lu, and Qingmin Liao "XSepConv: extremely separated convolution for efficient deep networks with large kernels", Proc. SPIE 11878, Thirteenth International Conference on Digital Image Processing (ICDIP 2021), 118780K (30 June 2021); https://doi.org/10.1117/12.2601043

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available