Saved in:
Bibliographic Details
Main Authors: Huang, Sining, Song, Yukun, Kang, Yixiao, Yu, Chang
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2409.14577
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914954747052032
author Huang, Sining
Song, Yukun
Kang, Yixiao
Yu, Chang
author_facet Huang, Sining
Song, Yukun
Kang, Yixiao
Yu, Chang
contents In the field of spatial computing, one of the most essential tasks is the pose estimation of 3D objects. While rigid transformations of arbitrary 3D objects are relatively hard to detect due to varying environment introducing factors like insufficient lighting or even occlusion, objects with pre-defined shapes are often easy to track, leveraging geometric constraints. Curved images, with flexible dimensions but a confined shape, are essential shapes often targeted in 3D tracking. Traditionally, proprietary algorithms often require specific curvature measures as the input along with the original flattened images to enable pose estimation for a single image target. In this paper, we propose a pipeline that can detect several logo images simultaneously and only requires the original images as the input, unlocking more effects in downstream fields such as Augmented Reality (AR).
format Preprint
id arxiv_https___arxiv_org_abs_2409_14577
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle AR Overlay: Training Image Pose Estimation on Curved Surface in a Synthetic Way
Huang, Sining
Song, Yukun
Kang, Yixiao
Yu, Chang
Computer Vision and Pattern Recognition
In the field of spatial computing, one of the most essential tasks is the pose estimation of 3D objects. While rigid transformations of arbitrary 3D objects are relatively hard to detect due to varying environment introducing factors like insufficient lighting or even occlusion, objects with pre-defined shapes are often easy to track, leveraging geometric constraints. Curved images, with flexible dimensions but a confined shape, are essential shapes often targeted in 3D tracking. Traditionally, proprietary algorithms often require specific curvature measures as the input along with the original flattened images to enable pose estimation for a single image target. In this paper, we propose a pipeline that can detect several logo images simultaneously and only requires the original images as the input, unlocking more effects in downstream fields such as Augmented Reality (AR).
title AR Overlay: Training Image Pose Estimation on Curved Surface in a Synthetic Way
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2409.14577