基于核化双线性卷积网络的细粒度图像分类
Kernelized Bilinear CNN Models for Fine-Grained Visual Recognition
查看参考文献30篇
文摘
|
双线性卷积网络( Bilinear CNN,B-CNN)在计算机视觉任务中有着广泛的应用. B-CNN通过对卷积层输出的特征进行外积操作,能够建模不同通道之间的线性相关,从而增强了卷积网络的表达能力.由于没有考虑特征图中通道之间的非线性关系,该方法无法充分利用通道之间所蕴含的更丰富信息.为了解决这一不足,本文提出了一种核化的双线性卷积网络,通过使用核函数的方式有效地建模特征图中通道之间的非线性关系,进一步增强卷积网络的表达能力.本文在三个常用的细粒度数据库CUB-200-2011、FGVC-Aircraft以及Cars上对本文方法进行了验证,实验表明本文方法在三个数据库上均优于同类方法. |
其他语种文摘
|
The bilinear convolutional neural network( B-CNN) has been widely used in computer vision. B-CNN can capture the linear correlation between different channels by performing the outer product operation on the features of the convolutional layer output, thus enhancing the representative ability of the convolutional network. Since the non-linear relationship between the channels in the feature map is not taken account of, this method cannot make full use of the richer information contained between the channels. In order to solve this problem, this paper proposes a kernelized bilinear convolutional neural network employing the kernel function to effectively capture the non-linear relationship between the channels in the feature map, and further enhancing the representative ability of the convolutional network. In this paper, the method is evaluated on three common fine-grained benchmarks CUB-200-2011,FGVC-Aircraft and Cars. Experiments show that our method is superior to its counterparts on all three benchmarks. |
来源
|
电子学报
,2019,47(10):2134-2141 【核心库】
|
DOI
|
10.3969/j.issn.0372-2112.2019.10.015
|
关键词
|
核化双线性聚合
;
双线性卷积网络
;
端到端学习
;
细粒度图像分类
|
地址
|
大连理工大学信息与通信工程学院, 辽宁, 大连, 116024
|
语种
|
中文 |
文献类型
|
研究性论文 |
ISSN
|
0372-2112 |
学科
|
自动化技术、计算机技术 |
基金
|
国家自然科学基金
|
文献收藏号
|
CSCD:6668666
|
参考文献 共
30
共2页
|
1.
Krizhevsky A. Imagenet classification with deep convolutional neural networks.
Advances in Neural Information Processing Systems,2012:1097-1105
|
CSCD被引
3083
次
|
|
|
|
2.
Deng J. Imagenet: A largescale hierarchical image database.
Proceedings of IEEE Conference on Computer Vision and Pattern Recognition,2009:248-255
|
CSCD被引
88
次
|
|
|
|
3.
柯圣财. 基于卷积神经网络和监督核哈希的图像检索方法.
电子学报,2017,45(1):157-163
|
CSCD被引
15
次
|
|
|
|
4.
王泽宇. 基于空间结构化推理深度融合网络的RGB-D场景解析.
电子学报,2018,46(5):1253-1258
|
CSCD被引
3
次
|
|
|
|
5.
李康. 基于卷积神经网络的鲁棒高精度目标跟踪算法.
电子学报,2018,46(9):2087-2093
|
CSCD被引
15
次
|
|
|
|
6.
邹承明. 基于多特征组合的细粒度图像分类方法.
计算机应用,2018,38(7):1853-1856,1861
|
CSCD被引
2
次
|
|
|
|
7.
Lin T Y. Bilinear CNN models for fine-grained visual recognition.
Proceedings of IEEE International Conference on Computer Vision,2015:1449-1457
|
CSCD被引
7
次
|
|
|
|
8.
Li P. Is second-order information helpful for large-scale visual recognition.
Proceedings of IEEE International Conference on Computer Vision,2017:2070-2078
|
CSCD被引
1
次
|
|
|
|
9.
Lin T Y. Improved bilinear pooling with CNNs.
British Machine Vision Conference,2017:1-12
|
CSCD被引
1
次
|
|
|
|
10.
Ioffe S. Batch normalization: Accelerating deep network training by reducing internal covariate shift.
International Conference on Machine Learning,2015:448-456
|
CSCD被引
216
次
|
|
|
|
11.
Maji S.
Fine-Grained Visual Classification of Aircraft,2013
|
CSCD被引
19
次
|
|
|
|
12.
Wah C.
The Caltech-Ucsd Birds-200-2011 Dataset. Technical report, Caltech,2011
|
CSCD被引
2
次
|
|
|
|
13.
Krause J. 3D object representations for fine-grained categorization.
Proceedings of IEEE International Conference on Computer Vision Workshops,2013:554-561
|
CSCD被引
6
次
|
|
|
|
14.
Gao Y. Compact bilinear pooling.
Proceedings of IEEE Conference on Computer Vision and attern Recognition,2016:317-326
|
CSCD被引
1
次
|
|
|
|
15.
Li Y. Factorized bilinear models for image recognition.
Proceedings of IEEE International Conference on Computer Vision,2017:2098-2106
|
CSCD被引
1
次
|
|
|
|
16.
Cui Y. Kernel pooling for convolutional neural networks.
Proceedings of IEEE Conference on Computer Vision and Pattern Recognition,2017:3049-3058
|
CSCD被引
1
次
|
|
|
|
17.
Wang Q. G2DeNet: Global Gaussian distribution embedding network and its application to visual recognition.
Proceedings of IEEE Conference on Computer Vision and Pattern Recognition,2017:2730-2739
|
CSCD被引
1
次
|
|
|
|
18.
Ionescu C.
Training deep networks with structured layers by matrix backpropagation,2015
|
CSCD被引
1
次
|
|
|
|
19.
Lin H T.
A Study on Sigmoid Kernels for SVM and the Training of Non-PSD Kernels by SMO-Type Methods. Technical Report,2003
|
CSCD被引
1
次
|
|
|
|
20.
Vedaldi A. Matconvnet: convolutional neural networks for matlab.
ACM International Conference on Multimedia,2015:689-692
|
CSCD被引
10
次
|
|
|
|
|