Salvato in:
Dettagli Bibliografici
Autori principali: Yan, Qingsong, Wang, Qiang, Zhao, Kaiyong, Chen, Jie, Li, Bo, Chu, Xiaowen, Deng, Fei
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2502.05859
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915144425013248
author Yan, Qingsong
Wang, Qiang
Zhao, Kaiyong
Chen, Jie
Li, Bo
Chu, Xiaowen
Deng, Fei
author_facet Yan, Qingsong
Wang, Qiang
Zhao, Kaiyong
Chen, Jie
Li, Bo
Chu, Xiaowen
Deng, Fei
contents Due to the rapid development of panorama cameras, the task of estimating panorama depth has attracted significant attention from the computer vision community, especially in applications such as robot sensing and autonomous driving. However, existing methods relying on different projection formats often encounter challenges, either struggling with distortion and discontinuity in the case of equirectangular, cubemap, and tangent projections, or experiencing a loss of texture details with the spherical projection. To tackle these concerns, we present SphereFusion, an end-to-end framework that combines the strengths of various projection methods. Specifically, SphereFusion initially employs 2D image convolution and mesh operations to extract two distinct types of features from the panorama image in both equirectangular and spherical projection domains. These features are then projected onto the spherical domain, where a gate fusion module selects the most reliable features for fusion. Finally, SphereFusion estimates panorama depth within the spherical domain. Meanwhile, SphereFusion employs a cache strategy to improve the efficiency of mesh operation. Extensive experiments on three public panorama datasets demonstrate that SphereFusion achieves competitive results with other state-of-the-art methods, while presenting the fastest inference speed at only 17 ms on a 512$\times$1024 panorama image.
format Preprint
id arxiv_https___arxiv_org_abs_2502_05859
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SphereFusion: Efficient Panorama Depth Estimation via Gated Fusion
Yan, Qingsong
Wang, Qiang
Zhao, Kaiyong
Chen, Jie
Li, Bo
Chu, Xiaowen
Deng, Fei
Computer Vision and Pattern Recognition
Due to the rapid development of panorama cameras, the task of estimating panorama depth has attracted significant attention from the computer vision community, especially in applications such as robot sensing and autonomous driving. However, existing methods relying on different projection formats often encounter challenges, either struggling with distortion and discontinuity in the case of equirectangular, cubemap, and tangent projections, or experiencing a loss of texture details with the spherical projection. To tackle these concerns, we present SphereFusion, an end-to-end framework that combines the strengths of various projection methods. Specifically, SphereFusion initially employs 2D image convolution and mesh operations to extract two distinct types of features from the panorama image in both equirectangular and spherical projection domains. These features are then projected onto the spherical domain, where a gate fusion module selects the most reliable features for fusion. Finally, SphereFusion estimates panorama depth within the spherical domain. Meanwhile, SphereFusion employs a cache strategy to improve the efficiency of mesh operation. Extensive experiments on three public panorama datasets demonstrate that SphereFusion achieves competitive results with other state-of-the-art methods, while presenting the fastest inference speed at only 17 ms on a 512$\times$1024 panorama image.
title SphereFusion: Efficient Panorama Depth Estimation via Gated Fusion
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2502.05859