Pub Date : 2024-06-12DOI: 10.1109/TMM.2024.3412388
Bo Wang;Fei Yu;Fei Wei;Yi Li;Wei Wang
Aremarkable number of backdoor attack methods have been proposed in the literature on deep neural networks (DNNs). However, it hasn't been sufficiently addressed in the existing methods of achieving true senseless backdoor attacks that are visually invisible and label-consistent. In this paper, we propose a new backdoor attack method where the labels of the backdoor images are perfectly aligned with their content, ensuring label consistency. Additionally, the backdoor trigger is meticulously designed, allowing the attack to evade DNN model checks and human inspection. Our approach employs an auto-encoder (AE) to conduct representation learning of benign images and interferes with salient classification features to increase the dependence of backdoor image classification on backdoor triggers. To ensure visual invisibility, we implement a method inspired by image steganography that embeds trigger patterns into the image using the DNN and enable sample-specific backdoor triggers. We conduct comprehensive experiments on multiple benchmark datasets and network architectures to verify the effectiveness of our proposed method under the metric of attack success rate and invisibility. The results also demonstrate satisfactory performance against a variety of defense methods.
{"title":"Invisible Intruders: Label-Consistent Backdoor Attack Using Re-Parameterized Noise Trigger","authors":"Bo Wang;Fei Yu;Fei Wei;Yi Li;Wei Wang","doi":"10.1109/TMM.2024.3412388","DOIUrl":"10.1109/TMM.2024.3412388","url":null,"abstract":"Aremarkable number of backdoor attack methods have been proposed in the literature on deep neural networks (DNNs). However, it hasn't been sufficiently addressed in the existing methods of achieving true senseless backdoor attacks that are visually invisible and label-consistent. In this paper, we propose a new backdoor attack method where the labels of the backdoor images are perfectly aligned with their content, ensuring label consistency. Additionally, the backdoor trigger is meticulously designed, allowing the attack to evade DNN model checks and human inspection. Our approach employs an auto-encoder (AE) to conduct representation learning of benign images and interferes with salient classification features to increase the dependence of backdoor image classification on backdoor triggers. To ensure visual invisibility, we implement a method inspired by image steganography that embeds trigger patterns into the image using the DNN and enable sample-specific backdoor triggers. We conduct comprehensive experiments on multiple benchmark datasets and network architectures to verify the effectiveness of our proposed method under the metric of attack success rate and invisibility. The results also demonstrate satisfactory performance against a variety of defense methods.","PeriodicalId":13273,"journal":{"name":"IEEE Transactions on Multimedia","volume":"26 ","pages":"10766-10778"},"PeriodicalIF":8.4,"publicationDate":"2024-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141968844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-06-11DOI: 10.1109/tmm.2024.3412330
Jiachen Kang, Wenjing Jia, Xiangjian He, Kin Man Lam
{"title":"Point Clouds Are Specialized Images: A Knowledge Transfer Approach for 3D Understanding","authors":"Jiachen Kang, Wenjing Jia, Xiangjian He, Kin Man Lam","doi":"10.1109/tmm.2024.3412330","DOIUrl":"https://doi.org/10.1109/tmm.2024.3412330","url":null,"abstract":"","PeriodicalId":13273,"journal":{"name":"IEEE Transactions on Multimedia","volume":"112 1","pages":""},"PeriodicalIF":7.3,"publicationDate":"2024-06-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141968846","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-06-11DOI: 10.1109/TMM.2024.3412389
Liang Zhang;Jiangwei Zhao;Qingbo Wu;Lili Pan;Hongliang Li
Unsupervised continual learning (UCL) has made remarkable progress over the past two years, significantly expanding the application of continual learning (CL). However, existing UCL approaches have only focused on transferring continual strategies from supervised to unsupervised. They have overlooked the relationship issue between visual features and representational continuity. This work draws attention to the texture bias problem in existing UCL methods. To address this problem, we propose a new UCL framework called InfoUCL, in which we develop InfoDrop contrastive loss to guide continual learners to extract more informative shape features of objects and discard useless texture features simultaneously. The proposed InfoDrop contrastive loss is general and can be combined with various UCL methods. Extensive experiments on various benchmarks have demonstrated that our InfoUCL framework can lead to higher classification accuracy and superior robustness to catastrophic forgetting.
{"title":"InfoUCL: Learning Informative Representations for Unsupervised Continual Learning","authors":"Liang Zhang;Jiangwei Zhao;Qingbo Wu;Lili Pan;Hongliang Li","doi":"10.1109/TMM.2024.3412389","DOIUrl":"10.1109/TMM.2024.3412389","url":null,"abstract":"Unsupervised continual learning (UCL) has made remarkable progress over the past two years, significantly expanding the application of continual learning (CL). However, existing UCL approaches have only focused on transferring continual strategies from supervised to unsupervised. They have overlooked the relationship issue between visual features and representational continuity. This work draws attention to the texture bias problem in existing UCL methods. To address this problem, we propose a new UCL framework called InfoUCL, in which we develop InfoDrop contrastive loss to guide continual learners to extract more informative shape features of objects and discard useless texture features simultaneously. The proposed InfoDrop contrastive loss is general and can be combined with various UCL methods. Extensive experiments on various benchmarks have demonstrated that our InfoUCL framework can lead to higher classification accuracy and superior robustness to catastrophic forgetting.","PeriodicalId":13273,"journal":{"name":"IEEE Transactions on Multimedia","volume":"26 ","pages":"10779-10791"},"PeriodicalIF":8.4,"publicationDate":"2024-06-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141968843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-06-06DOI: 10.1109/tmm.2024.3410672
Junyu Chen, Jie An, Hanjia Lyu, Christopher Kanan, Jiebo Luo
{"title":"Learning to Evaluate the Artness of AI-generated Images","authors":"Junyu Chen, Jie An, Hanjia Lyu, Christopher Kanan, Jiebo Luo","doi":"10.1109/tmm.2024.3410672","DOIUrl":"https://doi.org/10.1109/tmm.2024.3410672","url":null,"abstract":"","PeriodicalId":13273,"journal":{"name":"IEEE Transactions on Multimedia","volume":"22 1","pages":""},"PeriodicalIF":7.3,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141968849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-06-06DOI: 10.1109/tmm.2024.3410532
Mingsheng Li, Lin Zhang, Mingzhen Zhu, Zilong Huang, Gang Yu, Jiayuan Fan, Tao Chen
{"title":"Lightweight Model Pre-Training Via Language Guided Knowledge Distillation","authors":"Mingsheng Li, Lin Zhang, Mingzhen Zhu, Zilong Huang, Gang Yu, Jiayuan Fan, Tao Chen","doi":"10.1109/tmm.2024.3410532","DOIUrl":"https://doi.org/10.1109/tmm.2024.3410532","url":null,"abstract":"","PeriodicalId":13273,"journal":{"name":"IEEE Transactions on Multimedia","volume":"39 1","pages":""},"PeriodicalIF":7.3,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141968850","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-06-05DOI: 10.1109/tmm.2024.3410117
Lin Wang, Shiliang Sun, Jing Zhao
{"title":"VirPNet: A Multimodal Virtual Point Generation Network for 3D Object Detection","authors":"Lin Wang, Shiliang Sun, Jing Zhao","doi":"10.1109/tmm.2024.3410117","DOIUrl":"https://doi.org/10.1109/tmm.2024.3410117","url":null,"abstract":"","PeriodicalId":13273,"journal":{"name":"IEEE Transactions on Multimedia","volume":"74 1","pages":""},"PeriodicalIF":7.3,"publicationDate":"2024-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141968855","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：