Task-Driven Lens Design

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Yang, Xinge, Fu, Qiang, Nie, Yunfeng, Heidrich, Wolfgang
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915826028773376
author Yang, Xinge
Fu, Qiang
Nie, Yunfeng
Heidrich, Wolfgang
author_facet Yang, Xinge
Fu, Qiang
Nie, Yunfeng
Heidrich, Wolfgang
contents Classical lens design minimizes optical aberrations to produce sharp images, but is typically decoupled from downstream computer vision tasks. Existing end-to-end optical design learns optical encoding through joint optimization, but often suffers from an unstable training process. We propose task-driven lens design, a new optimization philosophy for joint optics-network systems. We freeze the pretrained vision model and optimize only the lens so that the image formation better fits the model's feature preferences. This network-frozen setting yields a low-dimensional and stable optimization process, enabling lens design from scratch without human intervention, thereby exploring a broader design space. Multiple computer vision experiments show that TaskLenses outperform classical ImagingLenses with the same or even fewer elements. Our analysis reveals that the learned optics exhibit long-tailed point spread functions, better preserving preferred structural cues when aberrations cannot be fully corrected. These results highlight task-driven design as a practical route for optical lenses that are compatible with modern vision models, and also inspire new optical design objectives beyond traditional aberration minimization.
format Preprint
id arxiv_https___arxiv_org_abs_2305_17185
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Task-Driven Lens Design
Yang, Xinge
Fu, Qiang
Nie, Yunfeng
Heidrich, Wolfgang
Computer Vision and Pattern Recognition
Graphics
Optics
Classical lens design minimizes optical aberrations to produce sharp images, but is typically decoupled from downstream computer vision tasks. Existing end-to-end optical design learns optical encoding through joint optimization, but often suffers from an unstable training process. We propose task-driven lens design, a new optimization philosophy for joint optics-network systems. We freeze the pretrained vision model and optimize only the lens so that the image formation better fits the model's feature preferences. This network-frozen setting yields a low-dimensional and stable optimization process, enabling lens design from scratch without human intervention, thereby exploring a broader design space. Multiple computer vision experiments show that TaskLenses outperform classical ImagingLenses with the same or even fewer elements. Our analysis reveals that the learned optics exhibit long-tailed point spread functions, better preserving preferred structural cues when aberrations cannot be fully corrected. These results highlight task-driven design as a practical route for optical lenses that are compatible with modern vision models, and also inspire new optical design objectives beyond traditional aberration minimization.
title Task-Driven Lens Design
topic Computer Vision and Pattern Recognition
Graphics
Optics
url https://arxiv.org/abs/2305.17185