Filters








470 Hits in 4.2 sec

Homography Estimation from Image Pairs with Hierarchical Convolutional Networks

Nathalie Japkowicz, Farzan Erlik Nowruzi, Robert Laganiere
2017 2017 IEEE International Conference on Computer Vision Workshops (ICCVW)  
In this paper, we introduce a hierarchy of twin convolutional regression networks to estimate the homography between a pair of images.  ...  At every convolutional network module, features from each image are extracted independently, given a shared set of kernels, also known as Siamese network model.  ...  Hierarchical Convolutional Network Model In this section, we present our network architecture in detail and show ho to calculate the four point homography estimate from an image pair.  ... 
doi:10.1109/iccvw.2017.111 dblp:conf/iccvw/JapkowiczNL17 fatcat:h4tijvg6t5fcdgsjjnicn423vi

Deep Homography Estimation and Its Application to Wall Maps of Wall-Climbing Robots

Qiang Zhou, Xin Li
2019 Applied Sciences  
We proposed HomographyFpnNet and obtained a smaller homography estimation error for a center-aligned image pair compared with the state of the art.  ...  In this study, we focused on the homography estimation between the camera image and wall map.  ...  [22] proposed a hierarchical model that is stacked by the twin convolutional regression networks to estimate the homography between a pair of images, and improved the prediction accuracy of four-point  ... 
doi:10.3390/app9142908 fatcat:mxernn75hfdn5j3k6zy2zje6me

STN-Homography: Direct Estimation of Homography Parameters for Image Pairs

Qiang Zhou, Xin Li
2019 Applied Sciences  
Estimating a 2D homography from a pair of images is a fundamental task in computer vision.  ...  Accordingly, we present STN-Homography, a neural network based on spatial transformer network (STN), to directly estimate the normalized homography matrix of an image pair.  ...  [24] used the hierarchy of twin convolutional regression networks to estimate the homography between a pair of images, and improved the prediction accuracy of four-point homography compared with that  ... 
doi:10.3390/app9235187 fatcat:zhvt7r37bvccdp4qshc27bexla

Deep Homography Estimation for Dynamic Scenes

Hoang Le, Feng Liu, Shu Zhang, Aseem Agarwala
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
However, these new methods do not consider dynamic content in input images. They train neural networks with only image pairs that can be perfectly aligned using homographies.  ...  Our experiments show that our method can robustly estimate homography for challenging scenarios with dynamic scenes, blur artifacts, or lack of textures.  ...  To deal with the large motion between two images, Nowruzi et al. developed a hierarchical neural network that iteratively refines homography estimation [11] .  ... 
doi:10.1109/cvpr42600.2020.00767 dblp:conf/cvpr/LeLZA20 fatcat:pql5rlpuxbckplf2yhuunyz25y

Deep Homography Estimation for Dynamic Scenes [article]

Hoang Le, Feng Liu, Shu Zhang, Aseem Agarwala
2020 arXiv   pre-print
However, these new methods do not consider dynamic content in input images. They train neural networks with only image pairs that can be perfectly aligned using homographies.  ...  Our experiments show that our method can robustly estimate homography for challenging scenarios with dynamic scenes, blur artifacts, or lack of textures.  ...  Fig. 1 (top), Fig. 11 , 12 (top), and Fig. 12 (bottom) originate from MS-COCO [19] , NUS [20] , Middleburry [2] , and Sintel [6] datasets respectively.  ... 
arXiv:2004.02132v1 fatcat:b3ftjjdvznavjlbivwa4bojp3q

Combining Convolutional Neural Network and Photometric Refinement for Accurate Homography Estimation

Lai Kang, Yingmei Wei, Yuxiang Xie, Jie Jiang, Yanming Guo
2019 IEEE Access  
pairs generated from a publicly available dataset.  ...  INDEX TERMS Homography estimation, convolutional neural networks (CNNs), photometric discrepancy, gradient-decent refinement.  ...  HCN-2 [25] stacks twin convolutional regression networks in a hierarchical way to estimate the homography between a pair of images.  ... 
doi:10.1109/access.2019.2933635 fatcat:weqyvzejnjccrdx5qz4p3cpcei

Homography estimation along short videos by recurrent convolutional regression network

Yang Mi, ,University of South Carolina, Columbia, 29208, USA, Kang Zheng, Song Wang
2020 Mathematical Foundations of Computing  
More specifically, we develop a recurrent convolutional regression network consisting of convolutional neural network (CNN) and recurrent neural network (RNN) with long short-term memory (LSTM) cells,followed  ...  by a regression layer for estimating the parameters of homography.  ...  Recently, many deep learning methods [12, 8, 40] were developed to learn the homography parameters from training image pairs with known homographies.  ... 
doi:10.3934/mfc.2020014 fatcat:yzfunsczjnbkdgh5kazcrbdp6u

Aggregation of convolutional neural network estimations of homographies by color transformations of the inputs

Miguel A. Molina-Cabello, David A. Elizondo, Rafael Marcos Luque-Baena, Ezequiel Lopez-Rubio
2020 IEEE Access  
It is rooted in a convolutional neural network for homography estimation, which is provided with a range of versions of the input pair of pictures.  ...  Each generated pair of images yields a distinct estimation of the homography, and then the estimations are combined together to obtain a final, more robust estimation.  ...  They would also like to thank the NVIDIA Corporation with the donation of two Titan X GPUs used for this research. They also acknowledge the grant of the Universidad de Málaga.  ... 
doi:10.1109/access.2020.2990744 fatcat:unpv4jxw6zbtxoyxepkw42lr7y

DFM: A Performance Baseline for Deep Feature Matching [article]

Ufuk Efe, Kutalmis Gokalp Ince, A. Aydin Alatan
2021 arXiv   pre-print
These estimates are simply based on dense matching of nearest neighbors at the terminal layer of VGG network outputs of the images to be matched.  ...  A novel image matching method is proposed that utilizes learned features extracted by an off-the-shelf deep neural network to obtain a promising performance.  ...  For this evaluation, four corners of the reference image is projected onto the pair image with estimated and ground-truth homographies, and the average pixel distance between estimated and ground-truth  ... 
arXiv:2106.07791v1 fatcat:z6gylowfkjffphbxwii77udnpy

Homography Estimation with Convolutional Neural Networks Under Conditions of Variance [article]

David Niblick, Avinash Kak
2020 arXiv   pre-print
to the estimation of homography.  ...  In this report, we analyze the performance of two recently published methods using Convolutional Neural Networks (CNNs) that are meant to replace the more traditional feature-matching based approaches  ...  Homography Estimation with Hierarchical Convolutional Networks (HH) The CNN-based approach for homography estimation as presented in [15] attempts to improve on the accuracy achievable by the previous  ... 
arXiv:2010.01041v2 fatcat:o7grfaqlgfblxm5uzei6nwuoba

Matching Disparate Image Pairs Using Shape-Aware ConvNets [article]

Shefali Srivastava, Abhimanyu Chopra, Arun CS Kumar, Suchendra M. Bhandarkar, Deepak Sharma
2018 arXiv   pre-print
The proposed correspondence determination scheme for matching disparate images exploits high-level shape cues that are derived from low-level local feature descriptors, thus combining the best of both  ...  An end-to-end trainable ConvNet architecture, that learns to harness the power of shape representation for matching disparate image pairs, is proposed.  ...  The MAE is computed using the homography matrix estimated by the network.  ... 
arXiv:1811.09889v1 fatcat:ztkgoyr6q5aj5dueb42xttnh7q

RANSAC-Flow: generic two-stage image alignment [article]

Xi Shen, François Darmon, Alexei A. Efros, Mathieu Aubry
2020 arXiv   pre-print
Fine alignment is learned in an unsupervised way by a deep network which optimizes a standard structural similarity metric (SSIM) between the two images, plus cycle-consistency.  ...  We propose a two-stage process: first, a feature-based parametric coarse alignment using one or more homographies, followed by non-parametric fine pixel-wise alignment.  ...  Acknowledgements: This work was supported by ANR project EnHerit ANR-17-CE23-0008, project Rapid Tabasco, NSF IIS-1633310, grants from SAP and Berkeley CLTC, and gifts from Adobe.  ... 
arXiv:2004.01526v2 fatcat:ltpp6c4gcnfmdo3pxdt2ulqqp4

SSORN: Self-Supervised Outlier Removal Network for Robust Homography Estimation [article]

Yi Li, Wenjie Pei, Zhenyu He
2022 arXiv   pre-print
Recent deep learning models intend to address the homography estimation problem using a single convolutional network.  ...  the traditional homography estimation pipeline.  ...  Currently, there are two research lines using deep convolutional networks (CNNs) for homography estimation.  ... 
arXiv:2208.14093v1 fatcat:fc2kblcyynahzaatkmtnxhl3jm

Deep Semantic Feature Matching Using Confidential Correspondence Consistency

Wei Lyu, Lang Chen, Zhong Zhou, Wei Wu
2020 IEEE Access  
This work aims to establish visual correspondences between a pair of images depicting objects of the same semantic category.  ...  Existing methods handle this task with handcrafted features, which cannot effectively fit the correlations between non-overlapping images.  ...  to the estimated homography.  ... 
doi:10.1109/access.2020.2966655 fatcat:a3srm3wiijgfvduchjtvy2gfti

Patch2Pix: Epipolar-Guided Pixel-Level Correspondences [article]

Qunjie Zhou, Torsten Sattler, Laura Leal-Taixe
2021 arXiv   pre-print
We show that our refinement network significantly improves the performance of correspondence networks on image matching, homography estimation, and localization tasks.  ...  Patch2Pix is weakly supervised to learn correspondences that are consistent with the epipolar geometry of an input image pair.  ...  Given a pair of images (I A , I B ), a CNN backbone with L layers extracts the feature maps from each image.  ... 
arXiv:2012.01909v3 fatcat:fs6jysjopbf3pmpo42mwrrtlxu
« Previous Showing results 1 — 15 out of 470 results