Image-Based Lifelogging: User Emotion Perspective

Bum, Junghyun; Choo, Hyunseung; Whang, Joyce Jiyoung

doi:10.32604/cmc.2021.014931

Cited by 2 publications

(2 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, we use the CelebA [29] and IMDB_crop, Wiki_crop [30] datasets as the training data for the federated learning face recognition gender classifcation task. Among them, the CelebA dataset was compiled and opened by the Chinese University of Hong Kong.…”

Section: Datasets and Experimental Confgurationmentioning

confidence: 99%

IPCADP-Equalizer: An Improved Multibalance Privacy Preservation Scheme against Backdoor Attacks in Federated Learning

Lian,

Zhang,

Chen

et al. 2023

International Journal of Intelligent Systems

View full text Add to dashboard Cite

Although there are some protection mechanisms in federated learning, its training process is still vulnerable to some powerful attacks, such as invisible backdoor attacks. Existing research work focuses more on how to prevent attacks in distributed training scenarios and improve the security of the FL training process, but it lacks consideration of utility and robustness, especially when the learning model of FL suffers from stealth backdoor attacks. This paper proposes an improved FL defense scheme IPCADP based on user-level differential privacy and variational autoencoders technology. The scheme can control and protect the privacy attribute of the image and can also eliminate the triggers that exist in the poisoned image. The experimental results show that compared with some existing defense schemes, IPCADP can defend against invisible backdoor attacks and improve the classification accuracy of the main task, while mitigating the impact of attacks on model robustness and stability. To a certain extent, the balance and unity of security, utility, and robustness are realized.

show abstract

Section: Datasets and Experimental Confgurationmentioning

confidence: 99%

IPCADP-Equalizer: An Improved Multibalance Privacy Preservation Scheme against Backdoor Attacks in Federated Learning

Lian,

Zhang,

Chen

et al. 2023

International Journal of Intelligent Systems

View full text Add to dashboard Cite

show abstract

“…Traditionally, image emotion recognition is grounded on statistics. The traditional models rely heavily on artificial visual features, which take lots of time and labor to construct, and incur a high cost of labeling the target dataset [12][13][14][15][16]. To solve the small sample problem of image emotion recognition, Narula et al [17] formulated a two-layer transfer CNN capable of extracting universal low-level image features and high-level semantic features, and thereby effectively solved the matching errors caused by the distribution difference between regions of interest (ROIs).…”

Section: Introductionmentioning

confidence: 99%

Online Learning Behavior Analysis Based on Image Emotion Recognition

Wang¹

2021

View full text Add to dashboard Cite

There are many limitations of the current online learning platforms in the monitoring of learning behaviors and the evaluation of teaching effects. For example, the platforms cannot sense and correct the changes in the learning states and emotions of the students. To overcome the limitations, this paper tries to analyze online learning behaviors based on image emotion recognition. Firstly, the flow of image emotion recognition was detailed to facilitate the analysis of online learning behaviors, and the key frames were extracted from human face images, using improved local binary pattern (LBP) and wavelet transform. Next, the authors constructed the structure for the system of online learning behavior analysis, proposed a learning emotion recognition method based on facial expressions, and established an image emotion classification model for online learning, based on the attention mechanism. Experimental results show that the proposed algorithm is effective in analyzing online learning behaviors based on image emotion recognition.

show abstract

Image-Based Lifelogging: User Emotion Perspective

Cited by 2 publications

References 20 publications

IPCADP-Equalizer: An Improved Multibalance Privacy Preservation Scheme against Backdoor Attacks in Federated Learning

IPCADP-Equalizer: An Improved Multibalance Privacy Preservation Scheme against Backdoor Attacks in Federated Learning

Online Learning Behavior Analysis Based on Image Emotion Recognition

Contact Info

Product

Resources

About