@article{ART003348218},
author={VO VAN PHAP and CHANG, HYOKYUNG},
title={PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data},
journal={ Journal of Software Forensics},
issn={3092-541X},
year={2026},
volume={22},
number={2},
pages={107-114},
doi={10.29056/jsf.2026.06.10}
TY - JOUR
AU - VO VAN PHAP
AU - CHANG, HYOKYUNG
TI - PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data
JO - Journal of Software Forensics
PY - 2026
VL - 22
IS - 2
PB - Korea Software Assessment and Valuation Society
SP - 107
EP - 114
SN - 3092-541X
AB - To tackle global model degradation and inter-client disparities under Non-IID data environments, we introduce PSR-FL, a novel federated learning method integrating a sparse representation-based feature learning structure with a lightweight personalization layer. By incorporating a k-sparse constraint to cap internal feature activations, our approach eliminates redundant computational loads while maintaining training stability. Furthermore, the personalization layer adapts directly to unique local data characteristics, minimizing performance variance without incurring additional communication overhead. Benchmarked on the RAPv2 and PETA datasets across diverse non-uniform settings, PSR-FL consistently outperforms conventional federated learning frameworks in both global accuracy and client stability. These empirical results prove its robust convergence and scalability, validating its readiness for distributed CCTV surveillance infrastructures.
KW - CCTV surveillance system;Federated learning;Personalized federated learning;Sparse representation;Non-IID data
DO - 10.29056/jsf.2026.06.10
ER -
VO VAN PHAP and CHANG, HYOKYUNG. (2026). PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data. Journal of Software Forensics, 22(2), 107-114.
VO VAN PHAP and CHANG, HYOKYUNG. 2026, "PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data", Journal of Software Forensics, vol.22, no.2 pp.107-114. Available from: doi:10.29056/jsf.2026.06.10
VO VAN PHAP, CHANG, HYOKYUNG "PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data" Journal of Software Forensics 22.2 pp.107-114 (2026) : 107.
VO VAN PHAP, CHANG, HYOKYUNG. PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data. 2026; 22(2), 107-114. Available from: doi:10.29056/jsf.2026.06.10
VO VAN PHAP and CHANG, HYOKYUNG. "PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data" Journal of Software Forensics 22, no.2 (2026) : 107-114.doi: 10.29056/jsf.2026.06.10
VO VAN PHAP; CHANG, HYOKYUNG. PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data. Journal of Software Forensics, 22(2), 107-114. doi: 10.29056/jsf.2026.06.10
VO VAN PHAP; CHANG, HYOKYUNG. PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data. Journal of Software Forensics. 2026; 22(2) 107-114. doi: 10.29056/jsf.2026.06.10
VO VAN PHAP, CHANG, HYOKYUNG. PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data. 2026; 22(2), 107-114. Available from: doi:10.29056/jsf.2026.06.10
VO VAN PHAP and CHANG, HYOKYUNG. "PSR-FL-based Learning Method for Enhancing CCTV Surveillance System Performance in Non-IID Data" Journal of Software Forensics 22, no.2 (2026) : 107-114.doi: 10.29056/jsf.2026.06.10
