RotateConv: Making Asymmetric Convolutional Kernels Rotatable

CASIA OpenIR > 智能感知与计算研究中心

	RotateConv: Making Asymmetric Convolutional Kernels Rotatable
	Ma JB(马佳彬)1,2 ; Guo WY(郭韦煜)1; Wang W(王威)1 ; Wang L(王亮)1
	2018-08
会议名称	International Conference on Pattern Recognition (ICPR)
会议日期	August 20-24 2018
会议地点	Beijing, China
摘要	In deep Convolutional Neural Networks(CNN), the design of kernel shapes influences a lot on the model size and performance. In this work, our proposed method, RotateConv, applies a novel kernel shape to massively reduce the number of parameters while maintaining considerable performance. The new shape is extremely simple as a line segment one, and we equip it with the rotatable ability which aims to learn diverse features with respect to different angles. The kernel weights and angles are learned simultaneously during end-to-end training via the standard back-propagation algorithm. There are two variants of RotateConv that only have 2 and 4 parameters respectively depending on whether using weight sharing, which are much compressed than the normal 3x3 kernel with 9 parameters. In experiments, we validate our RotateConv with two classical models, ResNet and DenseNet, on four image classification benchmark datasets, namely MNIST, CIFAR10, CIFAR100 and SVHN.
URL	查看原文
收录类别	EI
语种	英语
文献类型	会议论文
条目标识符	http://ir.ia.ac.cn/handle/173211/20885
专题	智能感知与计算研究中心
作者单位	1.中国科学院自动化研究所 2.中国科学院大学
第一作者单位	中国科学院自动化研究所
推荐引用方式 GB/T 7714	Ma JB,Guo WY,Wang W,et al. RotateConv: Making Asymmetric Convolutional Kernels Rotatable[C],2018.

条目包含的文件		下载所有文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
bare_conf.pdf（831KB）	会议论文		开放获取	CC BY-NC-SA	浏览下载