Saved in:
Bibliographic Details
Main Authors: Liu, Jingguo, Yu, Han, Li, Shigang, Li, Jianfeng
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.09216
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915385791479808
author Liu, Jingguo
Yu, Han
Li, Shigang
Li, Jianfeng
author_facet Liu, Jingguo
Yu, Han
Li, Shigang
Li, Jianfeng
contents Due to the current lack of large-scale datasets at the million-scale level, tasks involving panoramic images predominantly rely on existing two-dimensional pre-trained image benchmark models as backbone networks. However, these networks are not equipped to recognize the distortions and discontinuities inherent in panoramic images, which adversely affects their performance in such tasks. In this paper, we introduce a novel spherical sampling method for panoramic images that enables the direct utilization of existing pre-trained models developed for two-dimensional images. Our method employs spherical discrete sampling based on the weights of the pre-trained models, effectively mitigating distortions while achieving favorable initial training values. Additionally, we apply the proposed sampling method to panoramic image segmentation, utilizing features obtained from the spherical model as masks for specific channel attentions, which yields commendable results on commonly used indoor datasets, Stanford2D3D.
format Preprint
id arxiv_https___arxiv_org_abs_2507_09216
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle 360-Degree Full-view Image Segmentation by Spherical Convolution compatible with Large-scale Planar Pre-trained Models
Liu, Jingguo
Yu, Han
Li, Shigang
Li, Jianfeng
Computer Vision and Pattern Recognition
Due to the current lack of large-scale datasets at the million-scale level, tasks involving panoramic images predominantly rely on existing two-dimensional pre-trained image benchmark models as backbone networks. However, these networks are not equipped to recognize the distortions and discontinuities inherent in panoramic images, which adversely affects their performance in such tasks. In this paper, we introduce a novel spherical sampling method for panoramic images that enables the direct utilization of existing pre-trained models developed for two-dimensional images. Our method employs spherical discrete sampling based on the weights of the pre-trained models, effectively mitigating distortions while achieving favorable initial training values. Additionally, we apply the proposed sampling method to panoramic image segmentation, utilizing features obtained from the spherical model as masks for specific channel attentions, which yields commendable results on commonly used indoor datasets, Stanford2D3D.
title 360-Degree Full-view Image Segmentation by Spherical Convolution compatible with Large-scale Planar Pre-trained Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.09216