Filters








4,322 Hits in 4.4 sec

Multi-output Learning for Camera Relocalization

Abner Guzman-Rivera, Pushmeet Kohli, Ben Glocker, Jamie Shotton, Toby Sharp, Andrew Fitzgibbon, Shahram Izadi
2014 2014 IEEE Conference on Computer Vision and Pattern Recognition  
Furthermore, our method outperforms the stateof-the-art discriminative approach for camera relocalization.  ...  Thus, we propose and study methods for learning 'marginally relevant' predictors, and compare their performance when used with different selection procedures.  ...  Furthermore, our method outperforms state-of-the-art discriminative learning based methods for camera relocalization. Related Work on Multi-Output Prediction.  ... 
doi:10.1109/cvpr.2014.146 dblp:conf/cvpr/Guzman-RiveraKGSSFI14 fatcat:z4qkhpzjxvg3nmtptsmx366gzy

Visual Robot Relocalization Based on Multi-Task CNN and Image-Similarity Strategy

Tao Xie, Ke Wang, Ruifeng Li, Xinyue Tang
2020 Sensors  
The traditional CNN for 6D robot relocalization which outputs pose estimations does not interpret whether the model is making sensible predictions or just guessing at random.  ...  Thus, we propose a multi-task CNN for robot relocalization, which can simultaneously perform pose regression and scene recognition.  ...  Multi-Task Learning for Pose Regression and Scene Recognition The relocalization refers to the process to reproduce the location and orientation in three dimensions (namely, the camera pose of the images  ... 
doi:10.3390/s20236943 pmid:33291774 fatcat:djdilkxzovd55atnmk3xfwbcv4

Exploiting uncertainty in regression forests for accurate camera relocalization

Julien Valentin, Matthias Niebner, Jamie Shotton, Andrew Fitzgibbon, Shahram Izadi, Philip Torr
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Recent advances in camera relocalization use predictions from a regression forest to guide the camera pose optimization procedure.  ...  The main contributions of this work are (i) the extension of the state of the art on RGB-D camera relocalization by modeling and minimizing uncertainties for regression tree induction and predictions performed  ...  Multi-output learning for camera relocalization. In Computer Vision and Pattern Recognition.  ... 
doi:10.1109/cvpr.2015.7299069 dblp:conf/cvpr/ValentinNSFIT15 fatcat:rihvn3n4jjfebaihlccqif7wqq

Scene Coordinate Regression with Angle-Based Reprojection Loss for Camera Relocalization [article]

Xiaotian Li, Juha Ylioinas, Jakob Verbeek, Juho Kannala
2018 arXiv   pre-print
Image-based camera relocalization is an important problem in computer vision and robotics.  ...  The new loss also enables us to utilize available multi-view constraints, which further improve performance.  ...  Conclusion In this work, we have presented a new angle-based reprojection loss for learning scene coordinate regression for image-based camera relocalization.  ... 
arXiv:1808.04999v2 fatcat:462mrjhmyzeo5lz5jwuuulxatm

Hinted Networks [article]

Joel Lamy-Poirier, Anqi Xu
2018 arXiv   pre-print
We ground our investigations within the camera relocalization domain, and propose two variants, namely the Hinted Embedding and Hinted Residual networks, both applied to the PoseNet base model for regressing  ...  We present Hinted Networks: a collection of architectural transformations for improving the accuracies of neural network models for regression tasks, through the injection of a prior for the output prediction  ...  HINTED NETWORKS FOR CAMERA RELOCALIZATION A.  ... 
arXiv:1812.06297v1 fatcat:d426chvwqzbsvjqadurtxz5zju

Scene Coordinate Regression with Angle-Based Reprojection Loss for Camera Relocalization [chapter]

Xiaotian Li, Juha Ylioinas, Jakob Verbeek, Juho Kannala
2019 Lecture Notes in Computer Science  
Image-based camera relocalization is an important problem in computer vision and robotics.  ...  The new loss also enables us to utilize available multi-view constraints, which further improve performance.  ...  Conclusion In this work, we have presented a new angle-based reprojection loss for learning scene coordinate regression for image-based camera relocalization.  ... 
doi:10.1007/978-3-030-11015-4_19 fatcat:ntesaqmj6rhifdxj3au5ozv7qe

PoseNet: A Convolutional Network for Real-Time 6-DOF Camera Relocalization [article]

Alex Kendall, Matthew Grimes, Roberto Cipolla
2016 arXiv   pre-print
We present a robust and real-time monocular six degree of freedom relocalization system.  ...  This was made possible by leveraging transfer learning from large scale classification data.  ...  This dataset contains significant variation in camera height and was designed for RGB-D relocalization.  ... 
arXiv:1505.07427v4 fatcat:z5q36q3aezamhp2tis3o5lty24

Robust Neural Routing Through Space Partitions for Camera Relocalization in Dynamic Indoor Environments [article]

Siyan Dong, Qingnan Fan, He Wang, Ji Shi, Li Yi, Thomas Funkhouser, Baoquan Chen, Leonidas Guibas
2021 arXiv   pre-print
Our proposed algorithm is evaluated on the RIO-10 benchmark developed for camera relocalization in dynamic indoor environments.  ...  Localizing the camera in a known indoor environment is a key building block for scene mapping, robot navigation, AR, etc.  ...  Recently, deep learning features [1, 39, 38, 48] emerge for more robust and efficient relocalization.  ... 
arXiv:2012.04746v2 fatcat:y2ischdhune4pckkqng2pxvqea

PoseNet: A Convolutional Network for Real-Time 6-DOF Camera Relocalization

Alex Kendall, Matthew Grimes, Roberto Cipolla
2015 2015 IEEE International Conference on Computer Vision (ICCV)  
Relocalization results for an input image (top), the predicted camera pose of a visual reconstruction (middle), shown again overlaid in red on the original image (bottom).  ...  Figure 1 : PoseNet: Convolutional neural network monocular camera relocalization.  ...  This dataset contains significant variation in camera height and was designed for RGB-D relocalization.  ... 
doi:10.1109/iccv.2015.336 dblp:conf/iccv/KendallGC15 fatcat:upc5umehj5a6dapdtbsegejhbm

GN-Net: The Gauss-Newton Loss for Multi-Weather Relocalization [article]

Lukas von Stumberg, Patrick Wenzel, Qadeer Khan, Daniel Cremers
2019 arXiv   pre-print
Furthermore, we release an evaluation benchmark for relocalization tracking against different types of weather. Our benchmark is available at https://vision.in.tum.de/gn-net.  ...  To overcome this, we propose GN-Net: a network optimized with the novel Gauss-Newton loss for training weather invariant deep features, tailored for direct image alignment.  ...  Qualitative multi-weather evaluation Finally, we show a relocalization demo comparing our GN-Net to DSO.  ... 
arXiv:1904.11932v3 fatcat:wy62zn7ui5aonm5ysboltui6ba

Camera Relocalization by Exploiting Multi-View Constraints for Scene Coordinates Regression

Ming Cai, Huangying Zhan, Chamara Saroj Weerasekera, Kejie Li, Ian Reid
2019 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW)  
We propose a method for learning a scene coordinate regression model to perform accurate camera relocalization in a known environment from a single RGB image.  ...  For the warp error we explore both RGB values, and deep learned features, as the basis for the error.  ...  Conclusion We have proposed an efficient learning method for scene coordinate regression to carry out accurate 6DoF camera relocalization in a known scene from a single RGB image.  ... 
doi:10.1109/iccvw.2019.00469 dblp:conf/iccvw/CaiZWLR19 fatcat:dvtj5vmfuvgjzfaaeg4dzltjpe

Improved Visual Relocalization by Discovering Anchor Points [article]

Soham Saha, Girish Varma, C.V.Jawahar
2018 arXiv   pre-print
We address the visual relocalization problem of predicting the location and camera orientation or pose (6DOF) of the given input scene.  ...  Hence we propose a multi task loss function, which discovers the relevant anchor point, without needing the ground truth for it.  ...  The visual relocalization problem is defined as estimating the location as well as camera pose, given just the observed camera frame, without using any other sensor data.  ... 
arXiv:1811.04370v1 fatcat:pdrghhqg3nflzcnvglubvpmpka

Euler angles based loss function for camera relocalization with Deep learning [article]

Qiang Fang, Tianjiang Hu
2018 arXiv   pre-print
Deep learning has been applied to camera relocalization, in particular, PoseNet and its extended work are the convolutional neural networks which regress the camera pose from a single image.  ...  In this paper, we directly explore the three Euler angles as the orientation representation in the camera pose regressor.  ...  Acknowledgement This work was supported by the National Science Foundation for Young Scientists of China (Grant No:61703418) .  ... 
arXiv:1802.08851v1 fatcat:b3nkg3s67zhqpjkbwnia6f3za4

Near-field Perception for Low-Speed Vehicle Automation using Surround-view Fisheye Cameras [article]

Ciaran Eising, Jonathan Horgan, Senthil Yogamani
2021 arXiv   pre-print
Cameras are the primary sensor in automated driving systems. They provide high information density and are optimal for detecting road infrastructure cues laid out for human vision.  ...  Surround-view camera systems typically comprise of four fisheye cameras with 190+ field of view covering the entire 360 around the vehicle focused on near-field sensing.  ...  ACKNOWLEDGMENT We would like to thank our employer Valeo for encouraging advanced research.  ... 
arXiv:2103.17001v3 fatcat:uyy2ieomf5cajhened27n3n2mm

Modelling Uncertainty in Deep Learning for Camera Relocalization [article]

Alex Kendall, Roberto Cipolla
2016 arXiv   pre-print
We use a Bayesian convolutional neural network to regress the 6-DOF camera pose from a single RGB image.  ...  It obtains approximately 2m and 6 degrees accuracy for very large scale outdoor scenes and 0.5m and 10 degrees accuracy indoors.  ...  MODEL FOR DEEP REGRESSION OF CAMERA POSE For clarity, we briefly review PoseNet which we proposed in [8] to learn camera pose with deep regression.  ... 
arXiv:1509.05909v2 fatcat:zarj7j42s5f2rehlkoikfjx6ve
« Previous Showing results 1 — 15 out of 4,322 results