Pith. sign in

REVIEW 4 major objections 5 minor 62 references

Facial Expression Recognition with Controlled Privacy Preservation and Feature Compensation

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Facial identity can be removed from video expressions in the frequency domain while preserving most emotion recognition, yielding 78.84% accuracy and 2.02% identity leakage.

desk verdict Solid framework and coherent ablations, but the 2.01% privacy leakage claim is circular because the validator is the same classifier used to train the privacy enhancers. read the letter →

arxiv 2412.00277 v2 pith:UDE3QGLZ submitted 2024-11-29 cs.CV

Feng Xu , David Ahmedt-Aristizabal , Lars Petersson , Dadong Wang , Xun Li This is my paper · ORCID
classification cs.CV
keywords facialexpressionrecognitionprivacypreservationwavelettransformfeaturecompensationidentityleakagevideo-basedFERclosed-setfrequencydecomposition
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Video-based facial expression recognition (FER) faces a dilemma: hiding who the person is usually destroys the very facial dynamics that reveal emotion. This paper argues the dilemma can be split apart by frequency: identity lives mostly in low-frequency video components while expression dynamics live in high-frequency components. The proposed framework applies a wavelet transform, scrubs identity from each band separately with its own privacy enhancer steered by a frozen identity classifier, reconstructs the video, then uses a diffusion-based feature compensator guided by an expression classifier to restore expression detail. On the CREMA-D benchmark, the result is 78.84% expression accuracy with only 2.016% identity leakage, compared with 87.2% accuracy and 100% leakage without privacy. The claim is that decoupling privacy removal from utility learning, rather than jointly optimizing them, is what makes this trade-off possible.

What carries the argument

The load-bearing mechanism is the frequency split plus two independent controllers. A wavelet transform separates each video into low- and high-frequency sub-bands on the assumption that identity is primarily low-frequency and expression primarily high-frequency. Each band is passed through its own U-Net privacy enhancer ($F_{hpr}$ for high, $F_{lpr}$ for low) whose only training signal is to maximize the cross-entropy loss of a frozen identity budget controller (ResNet50, $C_{hpr}$/$C_{lpr}$) classifying the actor identity. The privacy-preserved bands are inverse-transformed into a reconstructed video, which is then passed through a feature compensator ($F_{fc}$, a U-ViT diffusion model) trained to minimize the cross-entropy loss of a frozen expression controller ($C_{fc}$, a ResNet50 pre-trained on DFEW), so that non-identity, expression-relevant features are enriched without re-introducing identity. A final frozen privacy leakage validator ($V_{pl}$, ResNet50 with the same weights as the controllers) converts identity recognition accuracy into the privacy leakage ratio, the paper's quantitative measure of privacy preservation.

What would settle it

Train an identity classifier on the high-frequency sub-bands of the original, unprivatized CREMA-D videos and measure its recognition accuracy: if it identifies the 91 actors well above chance, identity is present in high frequencies and the frequency-separation assumption is false for deep features. A complementary check is to run the full privacy pipeline with a hold-out identity classifier that was not used in training; if leakage rises well above 2%, the privacy guarantee is specific to the validator rather than to identity information itself.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that facial identity and facial expression are separable enough in the frequency domain that each can be processed independently: a wavelet transform splits each video into high- and low-frequency components, two U-Net privacy enhancers trained against a frozen ResNet50 identity classifier remove identity cues from both bands, and a U-ViT feature compensator trained against a frozen expression classifier rebuilds the expression-related detail that privacy enhancement strips away. The framework then measures what remains with a separate frozen identity classifier, the privacy leakage validator, which reports the proportion of frames whose true identity is still recognized. With this design, the paper reports near-chance identity leakage (2.016% versus a theoretical floor of about 1.099% for 91 identities) while keeping FER accuracy at 78.84%, within about 8 points of the no-privacy upper bound. The ablation study attributes the result to the per-frequency privacy enhancers, since leakage rises from 2.1% to 33.0% when they are removed, and to the controlled feature compensator, since accuracy falls from 78.8% to 48.7% when it is removed.

Load-bearing premise

The whole frequency-split design rests on the premise that a person's identity is carried almost entirely in low-frequency video components and expression in high-frequency components; if identity also survives in high-frequency details, or expression depends on low-frequency structure, the separation collapses and the privacy enhancers cannot remove identity without also removing expression.

Editorial extensions

If this is right

  • If the frequency-separation assumption holds for video faces, privacy enhancement can be applied per frequency band, giving finer control over the identity-utility trade-off than global blurring, which either leaks about 89% identity at comparable accuracy or drops accuracy to about 21% at comparable leakage.
  • Because the privacy modules and the FER classifier are trained separately, connected only by the feature compensator, the FER head can be swapped without retraining the privacy enhancers, and the same privacy pipeline could be reused for other downstream face-analysis tasks.
  • The privacy leakage ratio is an attribute-specific metric: with the appropriate controller and labels, the same validation mechanism can quantify leakage of other soft biometrics such as gender or skin colour.
  • Deployed systems should expect uneven per-expression performance: in the paper's results, sadness recognition falls from 80.8% to 40.9%, so applications serving particular emotions need per-class checks.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Not claimed by the paper: the frequency split is inherited from human psychophysics, but the authors do not verify it on deep features; a direct test is to train an identity classifier on high-frequency-only videos and check whether accuracy is well above chance.
  • Not claimed by the paper: the privacy guarantee is tied to the closed-set identity classifier used during training, so an attacker who re-identifies with a different face recognition model could report higher leakage than the 2.016% measured by the paper's own validator.
  • Not claimed by the paper: because the same controller supervises both frequency bands, the two enhancers are pushed to fool the same identity representation; separate controllers per band, or an adversarial loss that penalizes identity in the reconstructed video as a whole, might further close the gap to the roughly 1.1% chance floor.
  • Not claimed by the paper: the framework needs identity labels for training, so porting to in-the-wild video without identity annotations would require an unsupervised identity signal, such as clustering or self-supervised instance discrimination, as a proxy controller.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a two-stream framework for video-based facial expression recognition (FER) with facial identity privacy preservation. The original video is decomposed by a wavelet transform into low- and high-frequency components; each component is passed through a separate privacy enhancer (U-Net) trained against a frozen identity classifier (the 'identity budget controller'), then inverse-transformed, then enhanced by a U-ViT feature compensator trained against a frozen FER classifier, and finally evaluated by an R3D video FER classifier and a 'privacy leakage validator' (Vpl) that measures residual identity recognition. On CREMA-D, the framework reports 78.84% FER accuracy (versus 87.201% with no privacy preservation) and a 2.016% privacy leakage ratio (PLR), compared with baselines such as Gaussian blurring, a trade-off framework, optical flow, and face swapping. An ablation study and a U-Net-based recovery threat model are also reported.

Significance. The core idea—decoupling privacy removal from utility-task feature compensation and applying them separately in the frequency domain—is well motivated, and the engineering is coherent. The paper deserves credit for running a systematic ablation (Table 4) that isolates the contribution of each component; in particular, the feature compensator's effect is large and internally consistent (Task 5 at 48.656% versus 78.843% for the full framework). The threat model against image-recovery attacks is also a useful addition. However, the headline privacy claim rests on a circular evaluation: the privacy leakage validator uses the same pretrained identity classifier as the training adversary, so the reported 2.016% PLR only demonstrates that the privacy enhancers fool that one classifier. The frequency-domain separation assumption is also not verified on the model's own features or on CREMA-D. These issues must be addressed before the results can support the paper's central claim of identity concealment with preserved FER performance.

major comments (4)
  1. [Section 3.4 and Section 4.2] The privacy leakage ratio is circular. Algorithm 1 trains Fhpr and Flpr to maximize the cross-entropy loss of the frozen identity budget controllers Chpr and Clpr, and Section 3.4 states that Vpl is a ResNet50 with the same pretrained parameters as Chpr and Clpr. Therefore the 2.016% PLR measures how well the privacy enhancers fool the exact classifier they were optimized against, not whether identity is generally concealed. Since the closed-set test has 91 identities, chance accuracy is about 1.1%, so 2.016% is barely above chance for this specific model. The paper should report PLR under an independent face recognizer (different architecture and/or different training data), a linear probe on deep features of the reconstructed frames, and ideally an open-set identity-verification protocol; without at least one such independent measurement, the headline privacy claim is not supported.
  2. [Section 3.2 and Section 4.2] The load-bearing assumption that identity is primarily carried in low-frequency components and expression in high-frequency components is cited from psychophysical studies but never verified on the actual data used. The paper should provide quantitative evidence on CREMA-D, for example: train an identity classifier on the high-frequency-only components and an expression classifier on the low-frequency-only components, and report their accuracies; or measure how much identity information is recoverable from each band by a linear probe on the deep features of Fhpr and Flpr outputs. Without such a test, the claim that the two-stream frequency split is the reason for the privacy/utility trade-off remains an unsupported premise.
  3. [Table 1 and Table 4] All experimental numbers are reported as single-point accuracies with no error bars, no multiple runs, and no statistical significance testing. In Table 1, differences as small as 0.134 percentage points (78.843% vs. 77.778%) are used to compare methods, and in Table 4, task-level differences (e.g., Task 5 at 48.656% vs. Task 4 at 61.187%) may be consequential but could also reflect run-to-run variance. The paper should report means and standard deviations over at least three seeds, and, given that only CREMA-D is tested, ideally validate on a second closed-set video FER dataset before making general claims.
  4. [Section 4.5] The threat model evaluates only pixel-level image-recovery attacks with a U-Net. This does not address the central privacy threat for the paper's own PLR definition, which is identity classification by a potentially independent classifier. An adversary that simply trains a new identity classifier on the privacy-preserved videos (rather than attempting to reconstruct the original pixels) could recover identity at a much higher rate than 2.016%. The threat model should include an adaptive identity-classification attack using an independently trained model, or the paper should clearly scope its privacy claim to protection against the specific validator architecture.
minor comments (5)
  1. [Section 3.5 and Figure 2] The notation is inconsistent: Figure 2 labels the utility task as Fu, but Section 3.5 sometimes refers to Fu(Vi) and Fu(Ci) and other places call it the 'downstream task' without defining the symbol; please define all model symbols consistently in one place.
  2. [Table 3] The column headers use downward arrows for SSIM, PSNR, and PLR, but the interpretation differs: for privacy, lower SSIM/PSNR and lower PLR are better, while for image quality lower SSIM/PSNR would usually be worse. Please clarify the direction of desirability in the caption.
  3. [Section 4.1 and Section 4.2] The hyperparameters of the Gaussian blur baselines (the σ values used for 'Gaussian Blur 1' and 'Gaussian Blur 2') are not reported, making it impossible to reproduce the baseline results. Please provide the exact settings.
  4. [Section 4.6] In the ablation study, Task 1 removes the wavelet transform and inverse transform, but it is unclear whether the privacy enhancer then operates on the full-resolution video or on the combined frequency representation; please specify the exact input to Fhpr and Flpr in that task.
  5. [Abstract and Section 1] There are minor typographical issues, such as 'identify-related features' instead of 'identity-related features' and an extra parenthesis in 'the feature-rich video frames (Ci))'; please proofread the manuscript.

Circularity Check

1 steps flagged · score 6.0 of 10

Privacy leakage ratio is measured with the same frozen ResNet50 (Vpl) used as the privacy-enhancer adversary (Chpr/Clpr), so the headline 2.016% PLR largely reflects fitting to that validator rather than independent identity concealment.

  1. fitted input called prediction [Section 3.4 (Privacy Leakage Validation), with Algorithm 1 in Section 3.2 and Section 4.2]
    "Our privacy leakage validator Vpl uses ResNet50, pre-trained on the same dataset with identity labels as Chpr and Clpr, to recognize a random frame from the feature-compensated video. By maintaining consistency between Chpr, Clpr, and Vpl, we ensure that the privacy attribute being validated for leakage is the same one that was previously preserved."

    Algorithm 1 trains Fhpr and Flpr by maximizing the cross-entropy loss of yih=Chpr(Fhpr(Vih)) and yil=Clpr(Flpr(Vil)) against identity labels. Since Vpl is the same ResNet50 with the same pretrained parameters as Chpr and Clpr, the reported PLR (2.016%) is the accuracy of the very classifier the privacy enhancers were optimized to fool. The privacy claim is therefore the training objective renamed as an evaluation; an independent face recognizer or linear probe is needed to establish that identity is actually concealed.

full rationale

The FER accuracy is measured by an independent R3D network and the ablation study is internally coherent, so the paper is not entirely circular. However, the central privacy metric is not independent: the privacy leakage validator Vpl is the same network used as the frozen adversary during privacy-enhancer training. Thus the 2.016% PLR is essentially the outcome of optimizing against that validator, not evidence of general identity concealment. The threat model reuses the same PLR and therefore does not break the circularity. A different face recognizer or deep-feature probe is required to support the privacy claim. The author self-citations in the related work are not load-bearing for the main derivation.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The central claim rests on the frequency-separation assumption and on the assumption that fooling the training-time identity classifier transfers to unseen classifiers. Many implementation choices such as wavelet type, training lengths, and optimizer settings are not reported and act as hidden free parameters. No new physical entities are introduced.

free parameters (3)
  • Wavelet transform family and decomposition level = not reported
    The entire frequency-split pipeline depends on this choice, but the manuscript does not specify the wavelet type or level. A different choice could change PLR and ACC.
  • Privacy enhancer training duration and optimizer hyperparameters = not reported
    The degree of identity removal is controlled by training for 'epoch = 1 to N1' in Algorithm 1, but N1, learning rate, batch size, and stopping criterion are not given. These choices effectively set the privacy-utility operating point.
  • Baseline Gaussian blur sigma = tuned to 2.150% PLR and 77.778% ACC for the two baselines
    Gaussian Blur 1 and 2 sigmas are selected to match either the proposed PLR or ACC, making the baselines fair but not fixed a priori.
assumptions (6)
  • domain assumption Facial identity information is predominantly carried in low-frequency components, while facial expression information is predominantly carried in high-frequency components.
    Invoked in Section 3.2 and Section 4.2 as the basis for the wavelet split. The paper cites psychophysical studies [21, 8, 28], but does not verify the assumption on deep features or on CREMA-D video.
  • domain assumption Identity features removed by maximizing loss against a frozen identity classifier will also be removed for other face recognition models.
    The privacy leakage interpretation relies on this transferability. The paper only evaluates with Vpl, which shares weights with the training-time controllers, so transfer is assumed, not demonstrated.
  • domain assumption A U-ViT diffusion model pretrained on ImageNet can generate frames that restore expression features without reintroducing identity information.
    The feature compensator is trained only with an expression cross-entropy loss and no identity constraint. Its ImageNet priors are not shown to be identity-free on faces.
  • standard math The wavelet transform and inverse transform reconstruct the video frames without meaningful loss.
    Standard property of invertible wavelet transforms, invoked in Section 3.2. Not a risk.
  • domain assumption Closed-set identity labels on a 91-actor dataset are sufficient to validate privacy preservation.
    The paper acknowledges it only handles closed-set scenarios in Section 4.7. The privacy results may not transfer to open-set or in-the-wild identities.
  • domain assumption R3D pretrained on Kinetics-400 is an adequate utility model for video-based FER.
    Used as the expression classifier; prior work supports it, but no comparison with state-of-the-art FER architectures is made.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Facial Expression Recognition with Controlled Privacy Preservation and Feature Compensation." pith.science (2026). https://pith.science/paper/UDE3QGLZ

@misc{pith2026241200277,
  author       = {Pith},
  title        = {Pith review of: Facial Expression Recognition with Controlled Privacy Preservation and Feature Compensation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UDE3QGLZ}},
  note         = {Machine review of arXiv:2412.00277}
}
read the original abstract

Facial expression recognition (FER) systems raise significant privacy concerns due to the potential exposure of sensitive identity information. This paper presents a study on removing identity information while preserving FER capabilities. Drawing on the observation that low-frequency components predominantly contain identity information and high-frequency components capture expression, we propose a novel two-stream framework that applies privacy enhancement to each component separately. We introduce a controlled privacy enhancement mechanism to optimize performance and a feature compensator to enhance task-relevant features without compromising privacy. Furthermore, we propose a novel privacy-utility trade-off, providing a quantifiable measure of privacy preservation efficacy in closed-set FER tasks. Extensive experiments on the benchmark CREMA-D dataset demonstrate that our framework achieves 78.84% recognition accuracy with a privacy (facial identity) leakage ratio of only 2.01%, highlighting its potential for secure and reliable video-based FER applications.

Figures

Figures reproduced from arXiv: 2412.00277 by the authors.

Figure 1
Figure 1. Our framework: (a). Controlled privacy-preservation in the high- and low-frequencies (Freq.) with (b). Controlled fea￾ture compensation and (c). Privacy leakage validation. The results show the performance of different privacy preservation approaches with video-based FER on privacy leakage ratio and video FER ac￾curacy. GB 1 & 2, OF, Trade-off and IS are short for Gaussian Blurring 1 & 2, Optical Flow, Trade-off fra… view at source ↗
Figure 2
Figure 2. Illustration of our framework. 1 Controlled privacy-preservation: the original video (Vi) is transformed into high- and low￾frequency components (Vih and Vil), followed by privacy enhancement, which includes a privacy enhancer (Fhpr and Flpr) and its controller to strengthen privacy enhancement (Chpr and Clpr) for low- and high-frequency components respectively. The privacy-preserved high- and low-frequency frames a… view at source ↗
Figure 3
Figure 3. Fpr and Cpr denote the privacy enhancer and its con￾troller. They work on the original video. The privacy enhancer and identity budget controller (in purple) share identical architec￾tures and weights with Fhpr and Chpr, respectively. In Task 2, the controller-free privacy enhancer is trained while updating Fu. transformation steps in the controlled privacy preservation process. Without the separated privacy enhance… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: The structure of Tasks 4, 5 and 6. In Task 6, the param [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 57 canonical work pages

  1. [1]

    Health insurance portability and ac- countability act of 1996

    Accountability Act. Health insurance portability and ac- countability act of 1996. Public law, 104:191, 1996. 1

    work page 1996

  2. [2]

    A comprehensive re- view of facial expression recognition techniques.Multimedia Systems, 29(1):73–103, 2023

    R Rashmi Adyapady and B Annappa. A comprehensive re- view of facial expression recognition techniques.Multimedia Systems, 29(1):73–103, 2023. 3

    work page 2023

  3. [3]

    Deep facial analysis: A new phase i epilepsy evaluation using com- puter vision

    David Ahmedt-Aristizabal, Clinton Fookes, Kien Nguyen, Simon Denman, Sridha Sridharan, and Sasha Dionisio. Deep facial analysis: A new phase i epilepsy evaluation using com- puter vision. Epilepsy & Behavior, 82:17–24, 2018. 1

    work page 2018

  4. [4]

    A computer vision-based system for real-time detection of sleep onset in fatigued drivers

    Alexandra Branzan Albu, Ben Widsten, Tiange Wang, Julie Lan, and Jordana Mah. A computer vision-based system for real-time detection of sleep onset in fatigued drivers. In2008 IEEE intelligent vehicles symposium , pages 25–30. IEEE,

    work page

  5. [5]

    A compara- tive study on optical flow for facial expression analysis.Neu- rocomputing, 500:434–448, 2022

    Benjamin Allaert, Isaac Ronald Ward, Ioan Marius Bilasco, Chabane Djeraba, and Mohammed Bennamoun. A compara- tive study on optical flow for facial expression analysis.Neu- rocomputing, 500:434–448, 2022. 2

    work page 2022

  6. [6]

    Preserving privacy of face and facial expression in computer vision data collected in learning environments

    TS Ashwin and Ramkumar Rajendran. Preserving privacy of face and facial expression in computer vision data collected in learning environments. InInternational Conference on Ar- tificial Intelligence in Education , pages 561–567. Springer,

    work page

  7. [7]

    All are worth words: A vit backbone for diffusion models

    Fan Bao, Shen Nie, Kaiwen Xue, Yue Cao, Chongxuan Li, Hang Su, and Jun Zhu. All are worth words: A vit backbone for diffusion models. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 22669–22679, 2023. 3, 4, 6

    work page 2023

  8. [8]

    The independence of ex- pression and identity in face-processing: Evidence from neu- ropsychological case studies

    Sarah Bate and Rachel Bennetts. The independence of ex- pression and identity in face-processing: Evidence from neu- ropsychological case studies. Frontiers in psychology, 6:770,

    work page

Show all 62 references
  1. [9]

    A deep learning model for classifying human facial ex- pressions from infrared thermal images

    Ankan Bhattacharyya, Somnath Chatterjee, Shibaprasad Sen, Aleksandr Sinitca, Dmitrii Kaplun, and Ram Sarkar. A deep learning model for classifying human facial ex- pressions from infrared thermal images. Scientific reports, 11(1):20696, 2021. 2, 3

    2021

  2. [10]

    I know that person: Generative full body and face de-identification of people in images

    Karla Brkic, Ivan Sikiric, Tomislav Hrkac, and Zoran Kalafatic. I know that person: Generative full body and face de-identification of people in images. In 2017 IEEE Con- ference on Computer Vision and Pattern Recognition Work- shops (CVPRW), pages 1319–1328. IEEE, 2017. 2

    2017

  3. [11]

    Orientation- conditioned facial texture mapping for video-based facial re- mote photoplethysmography estimation

    Sam Cantrill, David Ahmedt-Aristizabal, Lars Petersson, Hanna Suominen, and Mohammad Ali Armin. Orientation- conditioned facial texture mapping for video-based facial re- mote photoplethysmography estimation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pat...

    2024

  4. [12]

    Crema-d: Crowd-sourced emotional multimodal actors dataset

    Houwei Cao, David G Cooper, Michael K Keutmann, Ruben C Gur, Ani Nenkova, and Ragini Verma. Crema-d: Crowd-sourced emotional multimodal actors dataset. IEEE transactions on affective computing, 5(4):377–390, 2014. 5

    2014

  5. [13]

    Moth-flame optimization based deep feature selection for facial expression recognition using thermal images

    Somnath Chatterjee, Debyarati Saha, Shibaprasad Sen, Diego Oliva, and Ram Sarkar. Moth-flame optimization based deep feature selection for facial expression recognition using thermal images. Multimedia Tools and Applications , 83(4):11299–11322, 2024. 2

    2024

  6. [14]

    Vgan- based image representation learning for privacy-preserving facial expression recognition

    Jiawei Chen, Janusz Konrad, and Prakash Ishwar. Vgan- based image representation learning for privacy-preserving facial expression recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition workshops, pages 1570–1579, 2018. 2, 3

    2018

  7. [15]

    Im- proving fashion landmark detection by dual attention fea- ture enhancement

    Ming Chen, Yingjie Qin, Lizhe Qi, and Yunquan Sun. Im- proving fashion landmark detection by dual attention fea- ture enhancement. In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision Workshops, pages 0–0, 2019. 3

    2019

  8. [16]

    Spact: Self-supervised privacy preservation for action recog- nition

    Ishan Rajendrakumar Dave, Chen Chen, and Mubarak Shah. Spact: Self-supervised privacy preservation for action recog- nition. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 20164–20173,

  9. [17]

    Vision-based fatigue driving recognition method integrating heart rate and facial features

    Guanglong Du, Tao Li, Chunquan Li, Peter X Liu, and Di Li. Vision-based fatigue driving recognition method integrating heart rate and facial features. IEEE transactions on intelli- gent transportation systems, 22(5):3089–3100, 2020. 1

    2020

  10. [18]

    Video-based emotion recognition using cnn-rnn and c3d hybrid networks

    Yin Fan, Xiangju Lu, Dian Li, and Yuanliu Liu. Video-based emotion recognition using cnn-rnn and c3d hybrid networks. In Proceedings of the 18th ACM international conference on multimodal interaction, pages 445–450, 2016. 3

    2016

  11. [19]

    Affective facial expressions recognition for human-robot interaction

    Diego R Faria, Mario Vieira, Fernanda CC Faria, and Cris- tiano Premebida. Affective facial expressions recognition for human-robot interaction. In 2017 26th IEEE international symposium on robot and human interactive communication (RO-MAN), pages 805–810. IEEE, 2017. 1

    2017

  12. [20]

    Ted-spad: Temporal distinctiveness for self- supervised privacy-preservation for video anomaly detec- tion

    Joseph Fioresi, Ishan Rajendrakumar Dave, and Mubarak Shah. Ted-spad: Temporal distinctiveness for self- supervised privacy-preservation for video anomaly detec- tion. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision, pages 13598–13609, 2023. 3

    2023

  13. [21]

    A comparison of spatial frequency tuning for the recognition of facial identity and facial expressions in adults and children

    Xiaoqing Gao and Daphne Maurer. A comparison of spatial frequency tuning for the recognition of facial identity and facial expressions in adults and children. Vision Research, 51(5):508–519, 2011. 2

    2011

  14. [22]

    Versatile audio-visual learning for handling single and multi modalities in emo- tion regression and classification tasks

    Lucas Goncalves, Seong-Gyun Leem, Wei-Cheng Lin, Berrak Sisman, and Carlos Busso. Versatile audio-visual learning for handling single and multi modalities in emo- tion regression and classification tasks. arXiv preprint arXiv:2305.07216, 2023. 5

    2023

  15. [23]

    Semi-supervised learning of multi-factor mod- els for face de-identification

    Ralph Gross, Latanya Sweeney, Fernando de la Torre, and Simon Baker. Semi-supervised learning of multi-factor mod- els for face de-identification. In 2008 IEEE Conference on Computer Vision and Pattern Recognition, pages 1–8, 2008. 2

    2008

  16. [24]

    Privacy-preserving face recognition with learn- able privacy budgets in frequency domain

    Jiazhen Ji, Huan Wang, Yuge Huang, Jiaxiang Wu, Xingkun Xu, Shouhong Ding, ShengChuan Zhang, Liujuan Cao, and Rongrong Ji. Privacy-preserving face recognition with learn- able privacy budgets in frequency domain. InEuropean Con- ference on Computer Vision, pages 475–491. Sprin...

    2022

  17. [25]

    Dfew: A 9 large-scale database for recognizing dynamic facial expres- sions in the wild

    Xingxun Jiang, Yuan Zong, Wenming Zheng, Chuangao Tang, Wanchuang Xia, Cheng Lu, and Jiateng Liu. Dfew: A 9 large-scale database for recognizing dynamic facial expres- sions in the wild. In Proceedings of the 28th ACM Interna- tional Conference on Multimedia , pages 2881–2889,...

    2020

  18. [26]

    The kinetics hu- man action video dataset

    Will Kay, Joao Carreira, Karen Simonyan, Brian Zhang, Chloe Hillier, Sudheendra Vijayanarasimhan, Fabio Viola, Tim Green, Trevor Back, Paul Natsev, et al. The kinetics hu- man action video dataset. arXiv preprint arXiv:1705.06950,

    arXiv

  19. [27]

    Factor- ized higher-order cnns with an application to spatio-temporal emotion estimation

    Jean Kossaifi, Antoine Toisoul, Adrian Bulat, Yannis Pana- gakis, Timothy M Hospedales, and Maja Pantic. Factor- ized higher-order cnns with an application to spatio-temporal emotion estimation. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recogni...

    2020

  20. [28]

    Emotion per- ception is mediated by spatial frequency content

    Devpriya Kumar and Narayanan Srinivasan. Emotion per- ception is mediated by spatial frequency content. Emotion, 11(5):1144, 2011. 2

    2011

  21. [29]

    De-identifying face image datasets while retaining facial expressions

    Andreas Leibl, Andreas Meißner, Stefan Altmann, Andreas Attenberger, and Helmut Mayer. De-identifying face image datasets while retaining facial expressions. In 2023 IEEE International Joint Conference on Biometrics (IJCB), pages 1–10. IEEE, 2023. 3

    2023

  22. [30]

    Intensity-aware loss for dynamic facial expression recogni- tion in the wild

    Hanting Li, Hongjing Niu, Zhaoqing Zhu, and Feng Zhao. Intensity-aware loss for dynamic facial expression recogni- tion in the wild. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, pages 67–75, 2023. 3

    2023

  23. [31]

    Deep facial expression recog- nition: A survey

    Shan Li and Weihong Deng. Deep facial expression recog- nition: A survey. IEEE transactions on affective computing, 13(3):1195–1215, 2020. 3

    2020

  24. [32]

    Color shift estimation-and-correction for image en- hancement

    Yiyu Li, Ke Xu, Gerhard Petrus Hancke, and Rynson WH Lau. Color shift estimation-and-correction for image en- hancement. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 25389– 25398, 2024. 2, 3

    2024

  25. [33]

    A facial ex- pression emotion recognition based human-robot interaction system

    Zhentao Liu, Min Wu, Weihua Cao, Luefeng Chen, Jianping Xu, Ri Zhang, Mengtian Zhou, and Junwei Mao. A facial ex- pression emotion recognition based human-robot interaction system. IEEE CAA J. Autom. Sinica, 4(4):668–676, 2017. 1

    2017

  26. [34]

    Do deepfakes adequately display emotions? a study on deepfake facial emotion expression

    Juan-Miguel L ´opez-Gil, Rosa Gil, and Roberto Garc ´ıa. Do deepfakes adequately display emotions? a study on deepfake facial emotion expression. Computational intelligence and neuroscience, 2022(1):1332122, 2022. 2, 7

    2022

  27. [35]

    Sgdr: Stochas- tic gradient descent with warm restarts

    Ilya Loshchilov and Frank Hutter. Sgdr: Stochas- tic gradient descent with warm restarts. arXiv preprint arXiv:1608.03983, 2016. 3

    2016 arXiv

  28. [36]

    Adverfacial: Privacy-preserving universal ad- versarial perturbation against facial micro-expression leak- ages

    Yin-Yin Low, Angeline Tanvy, Rapha ¨el C-W Phan, and Xi- aojun Chang. Adverfacial: Privacy-preserving universal ad- versarial perturbation against facial micro-expression leak- ages. In ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (...

    2022

  29. [37]

    On the reconstruction of face images from deep face tem- plates

    Guangcan Mai, Kai Cao, Pong C Yuen, and Anil K Jain. On the reconstruction of face images from deep face tem- plates. IEEE Trans Pattern Anal Mach Intell , 41(5):1188– 1202, 2018. 2

    2018

  30. [38]

    Privacy- preserving face recognition using random frequency compo- nents

    Yuxi Mi, Yuge Huang, Jiazhen Ji, Minyi Zhao, Jiaxiang Wu, Xingkun Xu, Shouhong Ding, and Shuigeng Zhou. Privacy- preserving face recognition using random frequency compo- nents. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision, pages 19673–19684, 2023. 2, 7

    2023

  31. [39]

    Privacy-preserving face recognition us- ing trainable feature subtraction

    Yuxi Mi, Zhizhou Zhong, Yuge Huang, Jiazhen Ji, Jian- qing Xu, Jun Wang, Shaoming Wang, Shouhong Ding, and Shuigeng Zhou. Privacy-preserving face recognition us- ing trainable feature subtraction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni...

    2024

  32. [40]

    An adversar- ial learning framework for preserving users’ anonymity in face-based emotion recognition

    Vansh Narula, Theodora Chaspari, et al. An adversar- ial learning framework for preserving users’ anonymity in face-based emotion recognition. arXiv preprint arXiv:2001.06103, 2020. 3

    2001 arXiv

  33. [41]

    Pre- serving privacy in image-based emotion recognition through user anonymization

    Vansh Narula, Kexin Feng, and Theodora Chaspari. Pre- serving privacy in image-based emotion recognition through user anonymization. In Proceedings of the 2020 Interna- tional Conference on Multimodal Interaction , pages 452– 460, 2020. 3

    2020

  34. [42]

    Connecting pixels to privacy and utility: Automatic redac- tion of private information in images

    Tribhuvanesh Orekondy, Mario Fritz, and Bernt Schiele. Connecting pixels to privacy and utility: Automatic redac- tion of private information in images. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recogni- tion, pages 8466–8475, 2018. 2

    2018

  35. [43]

    Stable hash generation for efficient privacy-preserving face identification

    Dail ´e Osorio-Roig, Christian Rathgeb, Pawel Drozdowski, and Christoph Busch. Stable hash generation for efficient privacy-preserving face identification. IEEE Transactions on Biometrics, Behavior, and Identity Science, 4(3):333–348,

  36. [44]

    Mobilenetv3: a deep learning technique for human face expressions iden- tification

    S Babu Rajendra Prasad and B Sai Chandana. Mobilenetv3: a deep learning technique for human face expressions iden- tification. International journal of information technology , 15(6):3229–3243, 2023. 2

    2023

  37. [45]

    Efficient privacy-preserving facial expression classifica- tion

    Yogachandran Rahulamathavan and Muttukrishnan Rajara- jan. Efficient privacy-preserving facial expression classifica- tion. IEEE Transactions on Dependable and Secure Com- puting, 14(3):326–338, 2015. 3

    2015

  38. [46]

    Black-box face recovery from identity features

    Anton Razzhigaev, Klim Kireev, Edgar Kaziakhmedov, Nurislam Tursynbek, and Aleksandr Petiushko. Black-box face recovery from identity features. In Computer Vision– ECCV 2020 Workshops: Glasgow, UK, August 23–28, 2020, Proceedings, Part V 16, pages 462–475. Springer, 2020. 2

    2020

  39. [47]

    U- net: Convolutional networks for biomedical image segmen- tation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- net: Convolutional networks for biomedical image segmen- tation. In Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, pa...

    2015

  40. [48]

    Comparing facial ex- pressions for face swapping evaluation with supervised con- trastive representation learning

    Felix Rosberg and Cristofer Englund. Comparing facial ex- pressions for face swapping evaluation with supervised con- trastive representation learning. In 2021 16th IEEE Interna- tional Conference on Automatic Face and Gesture Recogni- tion (FG 2021), pages 01–05. IEEE, 2021. 3

    2021

  41. [49]

    In search of a robust facial expressions recognition model: 10 A large-scale visual cross-corpus study

    Elena Ryumina, Denis Dresvyanskiy, and Alexey Karpov. In search of a robust facial expressions recognition model: 10 A large-scale visual cross-corpus study. Neurocomputing, 514:435–450, 2022. 5

    2022

  42. [50]

    Generative facial expressions and eye gaze behavior from prompts for multi-human-robot interaction

    Gabriel J Serfaty, Virgil O Barnard, and Joseph P Salisbury. Generative facial expressions and eye gaze behavior from prompts for multi-human-robot interaction. In Adjunct Pro- ceedings of the 36th Annual ACM Symposium on User Inter- face Software and Technology, pages 1–3, 2023. 1

    2023

  43. [51]

    Videoflow: Exploiting temporal cues for multi-frame optical flow estimation

    Xiaoyu Shi, Zhaoyang Huang, Weikang Bian, Dasong Li, Manyuan Zhang, Ka Chun Cheung, Simon See, Hongwei Qin, Jifeng Dai, and Hongsheng Li. Videoflow: Exploiting temporal cues for multi-frame optical flow estimation. In Proceedings of the IEEE/CVF International Conference on Com...

    2023

  44. [52]

    Detecting changes in facial temperature in- duced by a sudden auditory stimulus based on deep learning- assisted face tracking

    Saurabh Sonkusare, David Ahmedt-Aristizabal, Matthew J Aburn, Vinh Thai Nguyen, Tianji Pang, Sascha Frydman, Simon Denman, Clinton Fookes, Michael Breakspear, and Christine C Guo. Detecting changes in facial temperature in- duced by a sudden auditory stimulus based on deep lea...

    2019

  45. [53]

    A closer look at spatiotemporal convolutions for action recognition

    Du Tran, Heng Wang, Lorenzo Torresani, Jamie Ray, Yann LeCun, and Manohar Paluri. A closer look at spatiotemporal convolutions for action recognition. In Proceedings of the IEEE conference on Computer Vision and Pattern Recogni- tion, pages 6450–6459, 2018. 5

    2018

  46. [54]

    The eu general data protection regulation (gdpr)

    Paul V oigt and Axel V on dem Bussche. The eu general data protection regulation (gdpr). A Practical Guide, 1st Ed., Cham: Springer International Publishing, 10(3152676):10– 5555, 2017. 1

    2017

  47. [55]

    Rethinking the learning paradigm for dynamic facial expression recognition

    Hanyang Wang, Bo Li, Shuang Wu, Siyuan Shen, Feng Liu, Shouhong Ding, and Aimin Zhou. Rethinking the learning paradigm for dynamic facial expression recognition. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 17958–17968, 2023. 3

    2023

  48. [56]

    Privacy-preserving face recognition in the frequency domain

    Yinggui Wang, Jian Liu, Man Luo, Le Yang, and Li Wang. Privacy-preserving face recognition in the frequency domain. In Proceedings of the AAAI Conference on Artificial Intelli- gence, volume 36, pages 2558–2566, 2022. 1

    2022

  49. [57]

    Privacy-preserving deep action recogni- tion: An adversarial learning framework and a new dataset

    Zhenyu Wu, Haotao Wang, Zhaowen Wang, Hailin Jin, and Zhangyang Wang. Privacy-preserving deep action recogni- tion: An adversarial learning framework and a new dataset. IEEE Transactions on Pattern Analysis and Machine Intelli- gence, 44(4):2126–2139, 2020. 1, 3, 6

    2020

  50. [58]

    Mobilefaceswap: A lightweight framework for video face swapping

    Zhiliang Xu, Zhibin Hong, Changxing Ding, Zhen Zhu, Junyu Han, Jingtuo Liu, and Errui Ding. Mobilefaceswap: A lightweight framework for video face swapping. In Pro- ceedings of the AAAI Conference on Artificial Intelligence , volume 36, pages 2973–2981, 2022. 2, 3, 6

    2022

  51. [59]

    Low- resolution facial expression recognition: A filter learning perspective

    Yan Yan, Zizhao Zhang, Si Chen, and Hanzi Wang. Low- resolution facial expression recognition: A filter learning perspective. Signal Processing, 169:107370, 2020. 1

    2020

  52. [60]

    Undisturbed mental state assessment in the 5g era: a case study of depression detection based on facial expressions

    Minqiang Yang, Yu Ma, Zhenyu Liu, Hanshu Cai, Xiping Hu, and Bin Hu. Undisturbed mental state assessment in the 5g era: a case study of depression detection based on facial expressions. IEEE Wireless Communications, 28(3):46–53,

  53. [61]

    Altered fingerprints: Analysis and detection

    Soweon Yoon, Jianjiang Feng, and Anil K Jain. Altered fingerprints: Analysis and detection. IEEE transactions on pattern analysis and machine intelligence , 34(3):451–464,

  54. [62]

    Few-shot object counting with similarity-aware feature enhancement

    Zhiyuan You, Kai Yang, Wenhan Luo, Xin Lu, Lei Cui, and Xinyi Le. Few-shot object counting with similarity-aware feature enhancement. In Proceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision , pages 6315–6324, 2023. 3 11

    2023

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.