Learning from Less: Pose-Balanced Sampling for High-Performance Face Recognition
Sensors · 02.11.2026 · 26 (21) · 6937
DOI: 10.3390/s26216937 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 02.11.2026. Verfasst von Jhon I Pilataxi und 3 weiteren Autor:innen. Beteiligte Einrichtung: Escuela de Ingeniería Informática, Universidad de Valparaíso, Valparaíso 2362905, Chile. Hauptthemen laut MEDLINE: Automated Facial Recognition, Face. Die Arbeit nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Large-scale face recognition (FR) datasets are strongly skewed towards frontal images, which limits the ability of models to learn pose-robust representations and leaves frontal-profile verification as a persistent source of error. This paper proposes a pose-balanced sampling strategy that builds a compact training subset by conditioning selection on the per-identity head-pose distribution rather than on sample difficulty or redundancy. Yaw angles are estimated with 6DRepNet, and an equal number of frontal and profile images are retained for every identity. Applied to MS1Mv3, the strategy yields MS1Mv3-FP, a subset of 547,202 images (10.7% of the original) covering 78.5% of the identities. An SE-ResNet100 trained from scratch on MS1Mv3-FP using the ArcFace loss reaches up to 98.73%, 93.23%, and 94.74% verification accuracy on CFP-FP, CPLFW, and VGG2-FP (mean over three sampled subsets: 98.63%, 93.13%, and 94.50%), within 0.30, 0.70, and 0.86 percentage points (pp) of the same network trained on the complete dataset, while reducing training time from more than 100 h to approximately 11 h on a single GPU. Equal-size subsets built without pose balancing lose 0.8–3.6 pp on average, indicating that the pose composition of the subset, and not only its size, accounts for this result. Frontal–frontal verification accuracy is largely preserved, whereas age-gap verification degrades, delimiting the scope of the strategy. The resulting reduction in training time and data could support the development of pose-robust FR models for camera-based sensing systems operating under resource constraints.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Escuela de Ingeniería Informática, Universidad de Valparaíso, Valparaíso 2362905, Chile.
- Escuela de Informática y Telecomunicaciones, Universidad Diego Portales, Santiago 8370190, Chile.
- Departamento de Ingeniería de Sistemas y Computación, Universidad Católica del Norte, Antofagasta 1270709, Chile.
ORCID-Profile: Jhon I Pilataxi, Jorge E Zambrano, Robert Guamán-Rivera, Rodrigo Olivares
Zitieren
APA
Pilataxi, J. I., Zambrano, J. E., Guamán-Rivera, R., & Olivares, R. (2026). Learning from Less: Pose-Balanced Sampling for High-Performance Face Recognition. Sensors, 26(21), 6937. https://doi.org/10.3390/s26216937
Vancouver
Pilataxi JI, Zambrano JE, Guamán-Rivera R, Olivares R. Learning from Less: Pose-Balanced Sampling for High-Performance Face Recognition. Sensors. 2026;26(21):6937. doi:10.3390/s26216937
BibTeX
@article{Pilataxi2026,
author = {Pilataxi, Jhon I and Zambrano, Jorge E and Guamán-Rivera, Robert and Olivares, Rodrigo},
title = {Learning from Less: Pose-Balanced Sampling for High-Performance Face Recognition},
journal = {Sensors},
year = {2026},
volume = {26},
number = {21},
pages = {6937},
doi = {10.3390/s26216937}
}
EUVIMED