paper-with-me

홈 › Papers

Full explicit consistency constraints in uncalibrated multiple homography estimation

2018-05-07 · Wojciech Chojnacki, Zygmunt L. Szpak

We reveal a complete set of constraints that need to be imposed on a set of 3-by-3 matrices to ensure that the matrices represent genuine homographies associated with multiple planes between two views. We also show how to exploit the constraints to obtain more accurate estimates of homography matrices between two views. Our study resolves a long-standing research question and provides a fresh perspective and a more in-depth understanding of the multiple homography estimation task.

📄 PDF Abstract BibTeX arXiv:1805.02352

Code (0)

등록된 구현이 없습니다.

Tasks

Homography Estimation

Similar Papers 제목 키워드 기반

Unconstrained Multi-view Human Pose Estimation with Algebraic Priors

2026-04-27 · Xiaolin Qin, Qianlei Wang, Jiacen Liu, Chaoning Zhang 외 arxiv

Recovering 3D human pose from multi-view imagery typically relies on precise camera calibration, which is often unavailable in real-world scenarios, thereby severely limiting the applicability of existing methods. To ove…

Pose Estimation

Neural Architecture Search for Efficient Uncalibrated Deep Photometric Stereo

2021-10-11 · Francesco Sarno, Suryansh Kumar, Berk Kaya, Zhiwu Huang 외

We present an automated machine learning approach for uncalibrated photometric stereo (PS). Our work aims at discovering lightweight and computationally efficient PS neural networks with excellent surface normal accuracy…

Neural Architecture Search

Learning Precise 3D Manipulation from Multiple Uncalibrated Cameras

2020-02-21 · Iretiayo Akinola, Jacob Varley, Dmitry Kalashnikov

In this work, we present an effective multi-view approach to closed-loop end-to-end learning of precise manipulation tasks that are 3D in nature. Our method learns to accomplish these tasks using multiple statically plac…

Camera Calibration

Recollection from Pensieve: Novel View Synthesis via Learning from Uncalibrated Videos

2025-05-19 · Ruoyu Wang, Yi Ma, Shenghua Gao

Currently almost all state-of-the-art novel view synthesis and reconstruction models rely on calibrated cameras or additional geometric priors for training. These prerequisites significantly limit their applicability to …

3D geometryCamera Pose EstimationNovel View SynthesisPose Estimation

Unsupervised Learning of Depth and Ego-Motion from Monocular Video Using 3D Geometric Constraints

2018-02-15 · CVPR 2018 6 · Reza Mahjourian, Martin Wicke, Anelia Angelova

We present a novel approach for unsupervised learning of depth and ego-motion from monocular video. Unsupervised learning removes the need for separate supervisory signals (depth or ego-motion ground truth, or multi-view…

3D geometryDepth And Camera MotionDepth EstimationMonocular Depth Estimation+1