Filters








9 Hits in 2.4 sec

ActiveStereoNet: End-to-End Self-Supervised Learning for Active Stereo Systems [article]

Yinda Zhang, Sameh Khamis, Christoph Rhemann, Julien Valentin, Adarsh Kowdle, Vladimir Tankovich, Michael Schoenberg, Shahram Izadi, Thomas Funkhouser, Sean Fanello
2018 arXiv   pre-print
In this paper we present ActiveStereoNet, the first deep learning solution for active stereo systems.  ...  Due to the lack of ground truth, our method is fully self-supervised, yet it produces precise depth with a subpixel precision of 1/30th of a pixel; it does not suffer from the common over-smoothing issues  ...  This paper proposes the first end-to-end deep learning approach for active stereo that is trained fully self-supervised.  ... 
arXiv:1807.06009v1 fatcat:g4su4zk37jdpblr4xubxuihbii

ActiveStereoNet: End-to-End Self-supervised Learning for Active Stereo Systems [chapter]

Yinda Zhang, Sameh Khamis, Christoph Rhemann, Julien Valentin, Adarsh Kowdle, Vladimir Tankovich, Michael Schoenberg, Shahram Izadi, Thomas Funkhouser, Sean Fanello
2018 Lecture Notes in Computer Science  
In this paper we present ActiveStereoNet, the first deep learning solution for active stereo systems.  ...  Due to the lack of ground truth, our method is fully self-supervised, yet it produces precise depth with a subpixel precision of 1/30th of a pixel; it does not suffer from the common over-smoothing issues  ...  This paper proposes the first end-to-end deep learning approach for active stereo that is trained fully self-supervised.  ... 
doi:10.1007/978-3-030-01237-3_48 fatcat:wt3edayygjg5fd7jsk5xlu4cam

Self-Supervised Depth Completion for Active Stereo [article]

Frederik Warburg, Daniel Hernandez-Juarez, Juan Tarrio, Alexander Vakhitov, Ujwal Bonde, Pablo F. Alcantarilla
2022 arXiv   pre-print
These depth sensors, however, suffer from stereo artefacts and do not provide dense depth estimates.In this work, we present the first self-supervised depth completion method for active stereo systems  ...  Due to the nonexistence of publicly available active stereo datasets, we release a real dataset together with additional information for a publicly available synthetic dataset (TartanAir [42]) needed for  ...  CONCLUSIONS In this work we investigate self-supervised depth completion for active stereo systems.  ... 
arXiv:2110.03234v2 fatcat:mw2ekgmigzfbbhjdihuxn2j7mm

ActiveZero: Mixed Domain Learning for Active Stereovision with Zero Annotation [article]

Isabella Liu, Edward Yang, Jianyu Tao, Rui Chen, Xiaoshuai Zhang, Qing Ran, Zhu Liu, Hao Su
2021 arXiv   pre-print
In this paper we present a new framework, ActiveZero, which is a mixed domain learning solution for active stereovision systems that requires no real world depth annotation.  ...  Finally, we show how the method can be trained end-to-end and that each module is important for attaining the end result.  ...  This paper proposes an end-to-end learning stereo method that combines the advantages of self-supervised learning in the real domain and supervised learning in the simulation domain which we call mixed  ... 
arXiv:2112.02772v1 fatcat:cdnm4mmzdfab7jshdehzbpnmsu

Advances and Prospects of Vision-Based 3D Shape Measurement Methods

Guofeng Zhang, Shuming Yang, Pengyu Hu, Huiwen Deng
2022 Machines  
Specifically, the basic principles and typical techniques of triangulation-based measurement methods as well as their advantages and limitations are elaborated, and the learning-based techniques used for  ...  Vision-based three-dimensional (3D) shape measurement techniques have been widely applied over the past decades in numerous applications due to their characteristics of high precision, high efficiency  ...  [128] realized an end-to-end deep learning-based scheme for FPP systems.  ... 
doi:10.3390/machines10020124 fatcat:ws3pbwv6wzcole5jtca35c5fjm

SteReFo: Efficient Image Refocusing with Stereo Vision [article]

Benjamin Busam and Matthieu Hog and Steven McDonagh and Gregory Slabaugh
2019 arXiv   pre-print
To this end, high quality optical systems are used in DSLR cameras to focus on a specific depth plane while producing visually pleasing bokeh.  ...  It also enables computational video focus tracking for moving objects in addition to refocusing of static images.  ...  Its successor [56] extends the work with self-supervision to the domain of active sensing while maintaining the core efficiency. We build upon their work to leverage this computational advantage.  ... 
arXiv:1909.13395v1 fatcat:pygdnjn3pre3hj2dlynu5ftkim

SteReFo: Efficient Image Refocusing with Stereo Vision

Benjamin Busam, Matthieu Hog, Steven McDonagh, Gregory Slabaugh
2019 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW)  
To this end, high quality optical systems are used in DSLR cameras to focus on a specific depth plane while producing visually pleasing bokeh.  ...  It also enables computational video focus tracking for moving objects in addition to refocusing of static images.  ...  Its successor [56] extends the work with self-supervision to the domain of active sensing while maintaining the core efficiency. We build upon their work to leverage this computational advantage.  ... 
doi:10.1109/iccvw.2019.00411 dblp:conf/iccvw/BusamHMS19 fatcat:7oxfmmmgd5afjaaq4aztngcbye

HITNet: Hierarchical Iterative Tile Refinement Network for Real-time Stereo Matching [article]

Vladimir Tankovich, Christian Häne, Yinda Zhang, Adarsh Kowdle, Sean Fanello, Sofien Bouaziz
2021 arXiv   pre-print
At the time of writing, HITNet ranks 1st-3rd on all the metrics published on the ETH3D website for two view stereo, ranks 1st on most of the metrics among all the end-to-end learning approaches on Middlebury-v3  ...  This paper presents HITNet, a novel neural network architecture for real-time stereo matching.  ...  We would like to thank Shahram Izadi for support and enabling of this project.  ... 
arXiv:2007.12140v3 fatcat:nj63vbvjdjgbvc5oh6pluq7v5u

Depth Estimation by Combining Binocular Stereo and Monocular Structured-Light [article]

Yuhua Xu, Xiaoli Yang, Yushan Yu, Wei Jia, Zhaobi Chu, Yulan Guo
2022 arXiv   pre-print
It is well known that the passive stereo system cannot adapt well to weak texture objects, e.g., white walls. However, these weak texture targets are very common in indoor environments.  ...  In order to verify the effectiveness of the proposed system, we build a prototype and collect a test dataset in indoor scenes.  ...  ActiveStereoNet [47] is the first deep learning solution for active stereo systems. Due to the lack of ground truth, the network is designed to be fully self-supervised.  ... 
arXiv:2203.10493v1 fatcat:vj7rzbbewna2faw7ttfzozc5hm