2026

Procedia Computer Science

Rocznik: 2026 | Tom: 292

artykuł naukowy / referat

angielski

EN Focus breathing – the variation of effective field of view as a function of focus distance – introduces undesired scale changes in video sequences, particularly in photographic lenses lacking optomechanical compensation. Existing mitigation approaches are either hardware-dependent or unable to reliably distinguish breathing-induced scale changes from scene motion. This paper proposes a ray-tracing-based simulation framework for deterministic, scene-independent assessment of focus breathing in imaging lenses. The optical system is represented as a sequence of refractive surfaces; ray propagation is computed via Snell's law with aperture constraints enforced throughout. Image height at the sensor plane serves as a direct proxy for system magnification. A normalised focus ring position parameter (FRP ∈ [0, 1]) provides a lens-agnostic characterisation of the focusing range. For single internal focusing group lenses, high linearity of 𝑦𝑦!(𝐹𝐹𝐹𝐹𝐹𝐹) motivates a first-order approximation. A formal uncertainty model accounting for optical, numerical, and parametric error sources is derived. The framework is validated across seven lenses spanning four distinct focusing mechanisms.

7221 - 7230

10.1016/j.procs.2026.10.293

https://www.sciencedirect.com/science/article/pii/S1877050926039542

30th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES 2026), 9-11.09.2026, DUBLIN, Irlandia

CC BY-NC-ND (uznanie autorstwa - użycie niekomercyjne - bez utworów zależnych)

otwarte repozytorium

ostateczna wersja opublikowana

11.2026

w momencie opublikowania

5

70