EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Study of Loss Weight Balance in Lightweight Self-Distilled Crowd Counting

Muhammad Raza, Atta Ur Rahman, Pandula Pallewatta, Inayat Ur Rahman, Sahib Bahadar

Scientific Journal of Engineering Research · 29.05.2026 · 2 (3) · 409-420

DOI: 10.64539/sjer.v2i3.2026.493 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Scientific Journal of Engineering Research, online erschienen am 29.05.2026. Verfasst von Muhammad Raza und 4 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Lightweight crowd counting is important for real-time surveillance and resource-constrained deployment, where both computational efficiency and effective supervision are required. Although teacher-free self-distillation can improve lightweight density-regression models by guiding intermediate representations without an external teacher, the influence of composite loss weights in such frameworks has not been sufficiently analyzed. This paper presents a focused coefficient-wise loss-weight analysis within the Lightweight Self-Knowledge Distillation framework for single-image crowd counting. Instead of proposing a new architecture, the study investigates how the coefficients α, β, γ, and λ₂ affect optimization behavior and counting accuracy under a fixed experimental setup on ShanghaiTech Part B. Specifically, α controls intermediate feature alignment, β controls consistency supervision, γ controls direct density-regression supervision, and λ₂ controls the structural similarity term in the regression loss. The results show that moderate values of α and β improve performance by providing useful internal regularization, while excessive auxiliary weighting can slightly degrade accuracy. The analysis also indicates that γ should remain dominant because direct density-map regression is the primary learning signal. The best observed configuration is α = 6.0, β = 2.0, γ = 13.0, and λ₂ = 0.2, achieving 8.94 MAE and 11.51 RMSE on ShanghaiTech Part B. These findings highlight the importance of balanced supervision design within the evaluated LSKD framework on ShanghaiTech Part B.

Zitieren

APA

Raza, M., Rahman, A. U., Pallewatta, P., Rahman, I. U., & Bahadar, S. (2026). A Study of Loss Weight Balance in Lightweight Self-Distilled Crowd Counting. Scientific Journal of Engineering Research, 2(3), 409-420. https://doi.org/10.64539/sjer.v2i3.2026.493

Vancouver

Raza M, Rahman AU, Pallewatta P, Rahman IU, Bahadar S. A Study of Loss Weight Balance in Lightweight Self-Distilled Crowd Counting. Scientific Journal of Engineering Research. 2026;2(3):409-420. doi:10.64539/sjer.v2i3.2026.493

BibTeX
@article{Raza2026,
  author = {Raza, Muhammad and Rahman, Atta Ur and Pallewatta, Pandula and Rahman, Inayat Ur and Bahadar, Sahib},
  title = {A Study of Loss Weight Balance in Lightweight Self-Distilled Crowd Counting},
  journal = {Scientific Journal of Engineering Research},
  year = {2026},
  volume = {2},
  number = {3},
  pages = {409--420},
  doi = {10.64539/sjer.v2i3.2026.493}
}
Kostenlos starten →