Filters








430 Hits in 4.2 sec

Tanks and temples

Arno Knapitsch, Jaesik Park, Qian-Yi Zhou, Vladlen Koltun
2017 ACM Transactions on Graphics  
Ground-truth model for the Panther dataset, one of the datasets in the presented benchmark for large-scale scene reconstruction. ABSTRACT We present a benchmark for image-based 3D reconstruction.  ...  Additional Key Words and Phrases: Structure from motion, multi-view stereo, image-based reconstruction, large-scale scene reconstruction This is the authors' preprint of a SIGGRAPH 2017 paper.  ...  ACKNOWLEDGEMENTS We thank the anonymous reviewers for their detailed and professional reviews. Figures 1-4 were created using Mitsuba [Jakob 2010].  ... 
doi:10.1145/3072959.3073599 fatcat:kqqcacvhrzfiho4o2jradoopq4

AA-RMVSNet: Adaptive Aggregation Recurrent Multi-view Stereo Network [article]

Zizhuang Wei, Qingtian Zhu, Chen Min, Yisong Chen, Guoping Wang
2021 arXiv   pre-print
It ranks 1^st among all submissions on Tanks and Temples benchmark and achieves competitive results on DTU dataset, which exhibits strong generalizability and robustness.  ...  challenging regions, such as thin objects and large low-textured surfaces.  ...  Benchmarking on Tanks and Temples In order to evaluate the performance of our method under complex outdoor scenes, we test our method on Tanks and Temples benchmark as demonstrated in Tab. 2.  ... 
arXiv:2108.03824v1 fatcat:edkmh5kxwvboxegmj7chc2wosq

Recurrent MVSNet for High-Resolution Multi-View Stereo Depth Inference

Yao Yao, Zixin Luo, Shiwei Li, Tianwei Shen, Tian Fang, Long Quan
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
This reduces dramatically the memory consumption and makes high-resolution reconstruction feasible.  ...  Then, we further demonstrate the scalability of the proposed method on several large-scale scenarios, where previous learned approaches often fail due to the memory constraint.  ...  Some large scale reconstructions of Tanks and Temples dataset are shown in Fig. 5 .  ... 
doi:10.1109/cvpr.2019.00567 dblp:conf/cvpr/0008LLSFQ19 fatcat:6jckro7msfffpmqqzjcwlzoes4

Dense Hybrid Recurrent Multi-view Stereo Net with Dynamic Consistency Checking [article]

Jianfeng Yan, Zizhuang Wei, Hongwei Yi, Mingyu Ding, Runze Zhang, Yisong Chen, Guoping Wang, Yu-Wing Tai
2020 arXiv   pre-print
Our method ranks 1^st on the complex outdoor Tanks and Temples benchmark over all the methods.  ...  To further improve the accuracy and completeness of reconstructed point clouds, we leverage a dynamic consistency checking strategy instead of prefixed parameters and strategies widely adopted in existing  ...  Specifically, our method ranks 1 st in the complex large-scale outdoor Tanks and Temples benchmark over all existing methods.  ... 
arXiv:2007.10872v1 fatcat:m2qqs7ad75c63i5ud2v4ofulfe

Recurrent MVSNet for High-resolution Multi-view Stereo Depth Inference [article]

Yao Yao, Zixin Luo, Shiwei Li, Tianwei Shen, Tian Fang, Long Quan
2019 arXiv   pre-print
This reduces dramatically the memory consumption and makes high-resolution reconstruction feasible.  ...  Then, we further demonstrate the scalability of the proposed method on several large-scale scenarios, where previous learned approaches often fail due to the memory constraint.  ...  Some large scale reconstructions of Tanks and Temples dataset are shown in Fig. 5 .  ... 
arXiv:1902.10556v1 fatcat:dg4zo6fjxnexfhbph76ebzcvgu

DRI-MVSNet: A depth residual inference network for multi-view stereo images

Ying Li, Wenyue Li, Zhijie Zhao, JiaHao Fan, Sen Xiang
2022 PLoS ONE  
The results of extensive experiments show that DRI-MVSNet delivers competitive performance on the DTU and the Tanks & Temples datasets, and the accuracy and completeness of the point cloud reconstructed  ...  Due to the demand for large amounts of memory, prevalent methods of 3D reconstruction yield inaccurate results, because of which the highly accuracy reconstruction of a scene remains an outstanding challenge  ...  Tanks & Temples benchmark contains a large number of images of outdoor scenes.  ... 
doi:10.1371/journal.pone.0264721 pmid:35320265 pmcid:PMC8942269 fatcat:nyrzqymhwvcqtmzxatwsbqtn7q

Attention-Aware Multi-View Stereo

Keyang Luo, Tao Guan, Lili Ju, Yuesong Wang, Zhuo Chen, Yawei Luo
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Experimental results show that our approach achieves the best overall performance on the DTU dataset and the intermediate sequences of Tanks & Temples benchmark over many state-of-the-art MVS algorithms  ...  , we propose a novel attentionenhanced matching confidence volume, that combines the raw pixel-wise matching confidence from the extracted perceptual features with the contextual information of local scenes  ...  On the Tanks & Temples benchmark The Tanks & Temples is a widely used large-scale MVS benchmark and consists of two sequences: intermediate sequences and advanced sequences.  ... 
doi:10.1109/cvpr42600.2020.00166 dblp:conf/cvpr/LuoGJWCL20 fatcat:iqhvtwqy2fduhmlcg43oxt7k5m

HighRes-MVSNet: A Fast Multi-View Stereo Network for Dense 3D Reconstruction from High-Resolution Images

Rafael Weilharter, Friedrich Fraundorfer
2021 IEEE Access  
Although our method is aimed at much higher resolution images, we are still able to produce state-of-the-art results on the Tanks and Temples benchmark and achieve outstanding scores on the DTU benchmark  ...  INDEX TERMS Convolutional neural network, dense 3D reconstruction, multi-view stereo.  ...  Tanks and Temples.  ... 
doi:10.1109/access.2021.3050556 fatcat:ssupu7viubavdmij4fdgusenzq

BlendedMVS: A Large-Scale Dataset for Generalized Multi-View Stereo Networks

Yao Yao, Zixin Luo, Shiwei Li, Jingyang Zhang, Yufan Ren, Lei Zhou, Tian Fang, Long Quan
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
To create the dataset, we apply a 3D reconstruction pipeline to recover high-quality textured meshes from images of well-selected scenes.  ...  Compared with other computer vision tasks, it is rather difficult to collect a large-scale MVS dataset as it requires expensive active scanners and labor-intensive process to obtain ground truth 3D structures  ...  We thank Rui Chen for helping train and validate PointMVSNet [4] in our dataset.  ... 
doi:10.1109/cvpr42600.2020.00186 dblp:conf/cvpr/0008LLZRZFQ20 fatcat:olhoxgxurffwbiteo5j42jq2m4

Detail Preserved Surface Reconstruction from Point Cloud

Yang Zhou, Shuhan Shen, Zhanyi Hu
2019 Sensors  
In this paper, we put forward a new method for surface reconstruction from image-based point clouds.  ...  In particular, we introduce a new visibility model for each line of sight to preserve scene details without decreasing the noise filtering ability.  ...  However, in the case of large scale scenes, the computational and memory costs increase rapidly as scene scale grows. Consequently, this makes them unsuitable for large scale scene reconstruction.  ... 
doi:10.3390/s19061278 fatcat:7eip6w46qzekpixp43oqgzaynm

M^3VSNet: Unsupervised Multi-metric Multi-view Stereo Network [article]

Baichuan Huang, Hongwei Yi, Can Huang, Yijia He, Jingbin Liu, Xiao Liu
2020 arXiv   pre-print
powerful generalization ability on the Tanks and Temples benchmark with effective improvement.  ...  To improve the robustness and completeness of point cloud reconstruction, we propose a novel multi-metric loss function that combines pixel-wise and feature-wise loss function to learn the inherent constraints  ...  To evaluate the generalization ability of our proposed M 3 VSNet, we use the intermediate Tanks and Temples benchmark that has high-resolution images of outdoor large-scale scenes.  ... 
arXiv:2005.00363v2 fatcat:2v7bbzpgpbb2xmi24mcpw2b25e

BlendedMVS: A Large-scale Dataset for Generalized Multi-view Stereo Networks [article]

Yao Yao, Zixin Luo, Shiwei Li, Jingyang Zhang, Yufan Ren, Lei Zhou, Tian Fang, Long Quan
2020 arXiv   pre-print
To create the dataset, we apply a 3D reconstruction pipeline to recover high-quality textured meshes from images of well-selected scenes.  ...  Compared with other computer vision tasks, it is rather difficult to collect a large-scale MVS dataset as it requires expensive active scanners and labor-intensive process to obtain ground truth 3D structures  ...  By contrast, leaderboards of Tanks and Temples [14] and ETH3D [26] benchmarks are still dominated by classical MVS methods.  ... 
arXiv:1911.10127v2 fatcat:ibtqrzdfb5evnfjatshgiosmmy

M^3VSNet: Unsupervised Multi-metric Multi-view Stereo Network [article]

Baichuan Huang, Hongwei Yi, Can Huang, Yijia He, Jingbin Liu, Xiao Liu
2020 arXiv   pre-print
powerful generalization ability on the Tanks and Temples benchmark with effective improvement.  ...  To improve the robustness and completeness of point cloud reconstruction, we propose a novel multi-metric loss function that combines pixel-wise and feature-wise loss function to learn the inherent constraints  ...  To evaluate the generalization ability of our proposed M 3 VSNet, we use the intermediate Tanks and Temples benchmark that has high-resolution images of outdoor large-scale scenes.  ... 
arXiv:2004.09722v2 fatcat:dihmoxmhobbibeb5rs27j56x3e

PatchMatch-RL: Deep MVS with Pixelwise Depth, Normal, and Visibility [article]

Jae Yong Lee, Joseph DeGol, Chuhang Zou, Derek Hoiem
2021 arXiv   pre-print
We evaluate our method on widely used MVS benchmarks, ETH3D and Tanks and Temples (TnT), and compare to other state of the art learning based MVS models.  ...  However, non-learning based approaches still outperform for scenes with large depth ranges and sparser wide-baseline views, in part due to their PatchMatch optimization over pixelwise estimates of depth  ...  Tanks and Temples Benchmark With the same trained model, we evaluate on the Tanks and Temples [18] intermediate and advanced benchmarks which contain 8 intermediate and 6 advanced large-scale scenes  ... 
arXiv:2108.08943v1 fatcat:47vla23zoveifdaywzdokaxsoe

Multi-View Stereo Network with attention thin volume [article]

Zihang Wan
2022 arXiv   pre-print
Finally we follow the coarse to fine strategy and refine the depth sampling range scale by scale with the help of uncertainty estimation.  ...  With this approach, a more lightweight and efficient cost volume is constructed.  ...  The image resolution is 1600*1184, and the depth range of each scene is between 425mm to 935mm. Tank&Temples dataset contains two scene sets, namely intermediate and advanced.  ... 
arXiv:2110.08556v2 fatcat:jku5r42wdfbsbdlvrhyjgjdlli
« Previous Showing results 1 — 15 out of 430 results