EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Compressed-Domain Temporal Cues and Degradation-Aware Distillation for Lightweight Dynamic Facial Expression Recognition on Edge Devices

Alpamis Kutlimuratov, Islambek Saymanov, Bekmurod Khujamuratov, Dilfuza Uskenbaeva, Fazliddin Makhmudov, Mekhridin Rakhmanov, Ulugbek Bekmurodov, Jakhongir Karimboev, Kabul Khudaybergenov

Technologies · 05.11.2026 · 14 (11) · 729

DOI: 10.3390/technologies14110729 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Technologies, online erschienen am 05.11.2026. Verfasst von Alpamis Kutlimuratov und 8 weiteren Autor:innen. Die Arbeit nennt 64 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Dynamic facial expression recognition (DFER) has advanced rapidly, but leading systems use transformer backbones with more than 80 million parameters and assume fully decoded, high-quality video. Real deployments instead deliver small, heavily compressed and partially occluded faces to hardware with a few-watt power budget. We address this gap in three ways. First, DFEW-D and FERV39k-D form a reproducible degradation protocol that jointly sweeps face resolution, codec constant rate factor and occlusion over five severity levels and three codecs. Second, CoDA-Net is a 6.8 M-parameter network that reads motion vectors and residuals from the coded bitstream, combines them with appearance features from sparse intra frames and aggregates the sequence with a bidirectional state-space core, without reconstructing inter-coded frames. Third, quality-conditioned distillation uses a small quality head to estimate per-frame degradation and scales the supervision from a frozen teacher accordingly. We evaluate on DFEW, FERV39k and MAFW. On DFEW-D, CoDA-Net retains 84.2% of its reference-encoding weighted average recall at the most severe degradation level, against 66.7% for the strongest baseline, and runs at 32.3 clips/s on a Jetson Orin Nano and 5.4 clips/s on a Raspberry Pi 5. This suffices for stride-one prediction on the Jetson and for non-overlapping windows on the Pi 5. CoDA-Net is thus a robustness–efficiency trade-off for edge hardware, not a replacement for server-scale models.

Zitieren

APA

Kutlimuratov, A., Saymanov, I., Khujamuratov, B., Uskenbaeva, D., Makhmudov, F., Rakhmanov, M., Bekmurodov, U., Karimboev, J., & Khudaybergenov, K. (2026). Compressed-Domain Temporal Cues and Degradation-Aware Distillation for Lightweight Dynamic Facial Expression Recognition on Edge Devices. Technologies, 14(11), 729. https://doi.org/10.3390/technologies14110729

Vancouver

Kutlimuratov A, Saymanov I, Khujamuratov B, Uskenbaeva D, Makhmudov F, Rakhmanov M, et al. Compressed-Domain Temporal Cues and Degradation-Aware Distillation for Lightweight Dynamic Facial Expression Recognition on Edge Devices. Technologies. 2026;14(11):729. doi:10.3390/technologies14110729

BibTeX
@article{Kutlimuratov2026,
  author = {Kutlimuratov, Alpamis and Saymanov, Islambek and Khujamuratov, Bekmurod and Uskenbaeva, Dilfuza and Makhmudov, Fazliddin and Rakhmanov, Mekhridin and Bekmurodov, Ulugbek and Karimboev, Jakhongir and Khudaybergenov, Kabul},
  title = {Compressed-Domain Temporal Cues and Degradation-Aware Distillation for Lightweight Dynamic Facial Expression Recognition on Edge Devices},
  journal = {Technologies},
  year = {2026},
  volume = {14},
  number = {11},
  pages = {729},
  doi = {10.3390/technologies14110729}
}
Kostenlos starten →