Robust Image Classification with Cognitive-Driven Color Priors

doi:10.3390/electronics9111837

CASIA OpenIR > 多模态人工智能系统全国重点实验室 > 脑机融合与认知评估

	Robust Image Classification with Cognitive-Driven Color Priors
	Gu, Peng 1,2; Zhu, Chengfei2 ; Lan, Xiaosong2,3 ; Wang, Jie1,2 ; Li, Shuxiao1,2
发表期刊	ELECTRONICS
	2020-11-01
卷号	9 期号:11 页码:16
通讯作者	Li, Shuxiao(shuxiao.li@ia.ac.cn)
摘要	Existing image classification methods based on convolutional neural networks usually use a large number of samples to learn classification features hierarchically, causing the problems of over-fitting and error propagation layer by layer. Thus, they are vulnerable to adversarial samples generated by adding imperceptible disturbances to input samples. To address the above issue, we propose a cognitive-driven color prior model to memorize the color attributes of target samples inspired by the characteristics of human memory. At inference stage, color priors are indexed from the memory and fused with features of convolutional neural networks to achieve robust image classification. The proposed color prior model is cognitive-driven and has no training parameters, thus it has strong generalization and can effectively defend against adversarial samples. In addition, our method directly combines the features of the prior model with the classification probability of the convolutional neural network, without changing the network structure and its parameters of the existing algorithm. It can be combined with other adversarial attack defense methods, such as various preprocessing modules such as PixelDefense or adversarial training methods, to improve the robustness of image classification. Experiments on several benchmark datasets show that the proposed method improves the anti-interference ability of image classification algorithms.
关键词	adversarial samples color prior model image classifification
DOI	10.3390/electronics9111837
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[U19B2033] ; National Key RD Program[2019YFF0301801] ; Frontier Science and Technology Innovation Project[2019QY2404] ; Innovation Academy for Light-Duty Gas Turbine, Chinese Academy of Sciences[CXYJJ19-ZD-02]
项目资助者	National Natural Science Foundation of China ; National Key RD Program ; Frontier Science and Technology Innovation Project ; Innovation Academy for Light-Duty Gas Turbine, Chinese Academy of Sciences
WOS研究方向	Engineering
WOS类目	Engineering, Electrical & Electronic
WOS记录号	WOS:000592928700001
出版者	MDPI
引用统计
文献类型	期刊论文
条目标识符	http://ir.ia.ac.cn/handle/173211/41662
专题	多模态人工智能系统全国重点实验室_脑机融合与认知评估
通讯作者	Li, Shuxiao
作者单位	1.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China 2.Chinese Acad Sci, Inst Automat, Beijing 100190, Peoples R China 3.Chinese Acad Sci, Innovat Acad Light Duty Gas Turbine, Beijing 100080, Peoples R China
第一作者单位	中国科学院自动化研究所
通讯作者单位	中国科学院自动化研究所
推荐引用方式 GB/T 7714	Gu, Peng,Zhu, Chengfei,Lan, Xiaosong,et al. Robust Image Classification with Cognitive-Driven Color Priors[J]. ELECTRONICS,2020,9(11):16.
APA	Gu, Peng,Zhu, Chengfei,Lan, Xiaosong,Wang, Jie,&Li, Shuxiao.(2020).Robust Image Classification with Cognitive-Driven Color Priors.ELECTRONICS,9(11),16.
MLA	Gu, Peng,et al."Robust Image Classification with Cognitive-Driven Color Priors".ELECTRONICS 9.11(2020):16.