International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 11 | Nov 2025
p-ISSN: 2395-0072
www.irjet.net
De-identification Meets Emotion: A Review of Privacy-Enhanced Facial Expression Recognition Frameworks FATHIMATH SUHAIMA PU MSc Computer Science Student, St. Thomas (Autonomous) College, Thrissur 680001, Kerala, India ---------------------------------------------------------------------***--------------------------------------------------------------------To address these concerns, the research community has Abstract - Facial Expression Recognition (FER) has become
begun exploring privacy-preserving FER (PP-FER) frameworks that can maintain recognition performance while suppressing identity information. A recent advancement by Xu et al. introduced a frequency-based dualstream approach that decomposes video frames into lowfrequency (identity) and high-frequency (expression) components using wavelet transforms. Their method applies controlled privacy enhancement to each component, followed by a feature compensator that restores expression features compromised during anonymization. With a 78.84% accuracy and only 2.01% identity leakage on the CREMA-D dataset, the approach offers a compelling balance between privacy and utility.
increasingly vital in emotion-aware systems such as intelligent tutoring, healthcare, and surveillance. However, the processing of facial video data raises significant privacy concerns due to the inherent exposure of biometric identity information. This review paper explores recent advancements in privacypreserving FER, with a focus on federated learning, frequencybased de-identification,and expression-preserving face anonymization techniques. Central to this discussion is the novel dual-frequency framework proposed by Xu et al., which decomposes video data into high- and low-frequency components to isolate expression and identity features, respectively. By applying controlled privacy enhancement and feature compensation, the method achieves low identity leakage (2.01%) with high FER accuracy (78.84%) on the CREMA-D dataset.
Complementary to this, multiple other privacy-aware FER strategies have emerged in the literature. The RAPOO system enables mobile-based FER via crowdsensing while protecting privacy at the source. Meanwhile, GAN-based face anonymization techniques retain expression cues while removing biometric features in educational and public settings. Other models focus on enhancing expression recognition in the presence of privacy constraints using hybrid CNN–RNN networks and advanced temporal modeling. These innovations show that integrating FER with privacy-preserving mechanisms is not only feasible but increasingly essential in real-world deployments.
Complementing this, recent literature demonstrates parallel efforts to balance privacy and utility. For instance, RAPOO utilizes mobile crowdsensing and lightweight encryption to enable privacy-aware FER on edge devices. StyleGAN-based image anonymization methods retain expressive features while effectively masking identities in educational and generalpurpose FER datasets. Additionally, hybrid CNN–ConvLSTM architectures and compound emotion models further improve spatiotemporal understanding and classification robustness. Across these approaches, a common challenge remains: designing architectures that maintain recognition accuracy without compromising personal privacy.
Applications of Facial Emotion Recognition-
Key Words: Facial Expression Recognition (FER); Privacy Preservation; Wavelet Transform; De-identification; Feature Compensation; Identity Leakage; Video-based Emotion Recognition; Deep Learning.
I.
FER is a significant field because facial expressions are a crucial form of non-verbal communication, conveying up to 55% of emotional information. The ability to automatically recognize these expressions has led to a wide range of applications:
INTRODUCTION
Facial Expression Recognition (FER) plays a crucial role in enabling machines to interpret human emotions through facial cues. It has widespread applications in diverse domains such as mental health monitoring, driver fatigue detection, intelligent tutoring systems, and human–robot interaction. With the growing use of video-based FER in these sensitive environments, preserving the privacy of users—especially their facial identity—has become a pressing concern. FER systems typically rely on large volumes of facial video data, which inherently contain personally identifiable information. This raises ethical, legal, and societal questions surrounding data protection, user consent, and potential misuse.
© 2025, IRJET
|
Impact Factor value: 8.315
|
Healthcare: FER systems are used to detect conditions like autism and neurodegenerative diseases and can help predict psychotic disorders or depression. In elderly care, it can be used to monitor patients and identify those who need assistance, including for suicide prevention.
Education: In educational settings, FER can be used for intelligent tutoring and to monitor student moods and attention levels, providing insights into engagement and comprehension.
ISO 9001:2008 Certified Journal
|
Page 559