☆ 3.8 Proceedings Paper

MVSTER: Epipolar Transformer for Efficient Multi-view Stereo

COMPUTER VISION, ECCV 2022, PT XXXI (2022)

Related references

Note: Only part of the references are listed.

Proceedings Paper Computer Science, Artificial Intelligence

Tracking Objects as Pixel-Wise Distributions

Zelin Zhao et al.

Summary: This paper proposes a transformer-based architecture called P3AFormer for multi-object tracking. Objects are tracked as pixel-wise distributions, and the architecture utilizes flow information to guide the propagation of pixel-wise features. P3AFormer achieves state-of-the-art performance on multiple benchmark datasets.

COMPUTER VISION, ECCV 2022, PT XXII (2022)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo

Fangjinhua Wang et al.

Summary: IterMVS is a new data-driven method for high-resolution multi-view stereo. It encodes and refines the pixel-wise probability distributions of depth using a GRU-based estimator, and combines traditional classification and regression for depth map extraction. The method has been validated for efficiency and effectiveness, and compared with state-of-the-art methods.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) (2022)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

MVS2D: Efficient Multi-view Stereo via Attention-Driven 2D Convolutions

Zhenpei Yang et al.

Summary: MVS2D is a highly efficient multi-view stereo algorithm that seamlessly integrates multi-view constraints into single-view networks via an attention mechanism. It is at least 2 times faster than all notable counterparts and achieves precise depth estimations and 3D reconstructions, achieving state-of-the-art results.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) (2022)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Rethinking Depth Estimation for Multi-View Stereo: A Unified Representation

Rui Peng et al.

Summary: This paper proposes a novel representation method for depth estimation, called Unification, which combines the advantages of regression and classification. A new loss function is designed to address the challenge of sample imbalance. Experimental results show that our model outperforms other methods on different datasets and demonstrates the best generalization ability.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) (2022)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

TransMVSNet: Global Context-aware Multi-view Stereo Network with Transformers

Yikang Ding et al.

Summary: This paper introduces TransMVSNet, which leverages feature matching Transformer and other techniques to achieve state-of-the-art performance in multi-view stereo matching.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) (2022)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

AA-RMVSNet: Adaptive Aggregation Recurrent Multi-view Stereo Network

Zizhuang Wei et al.

Summary: The paper presents a novel recurrent multi-view stereo network, AA-RMVSNet, based on LSTM with adaptive aggregation modules for enhancing 3D reconstruction accuracy and completeness. The lightweight and effective adaptive aggregation modules improve the performance on challenging regions and varying occlusion, while the hybrid network structure enables high-resolution reconstruction and finer hypothetical plane sweep. The end-to-end trained network achieves excellent performance on various datasets, ranking 1st on Tanks and Temples benchmark and producing competitive results on DTU dataset, showing strong generalizability and robustness.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

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

Jae Yong Lee et al.

Summary: In this paper, an end-to-end trainable PatchMatch-based MVS approach is proposed, combining the advantages of trainable costs and regularizations with pixelwise estimates. Through reinforcement learning, the non-differentiable PatchMatch optimization is optimized, achieving satisfactory results.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Revisiting Stereo Depth Estimation From a Sequence-to-Sequence Perspective with Transformers

Zhaoshuo Li et al.

Summary: This research utilizes a sequence-to-sequence correspondence perspective to replace cost volume construction, achieves promising results, and demonstrates generalization across different domains without fine-tuning.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

EPP-MVSNet: Epipolar-assembling based Depth Prediction for Multi-view Stereo

Xinjun Ma et al.

Summary: EPP-MVSNet is a novel deep learning network for 3D reconstruction from multi-view stereo, which achieves effective and efficient 3D construction by optimizing feature aggregation, introducing epipolar-based kernel, and entropy-based refining strategy.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

The Temporal Opportunist: Self-Supervised Multi-Frame Monocular Depth

Jamie Watson et al.

Summary: The study introduces a new self-supervised monocular depth estimation method called ManyDepth, which can utilize sequence information at test time, and experiments show that it outperforms existing self-supervised baselines.

2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021 (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

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

Vladimir Tankovich et al.

Summary: HITNet is a novel neural network architecture for real-time stereo matching that achieves high accuracy with efficient computation. Its approach of using multi-resolution initialization, differentiable 2D geometric propagation, and warping mechanisms for disparity inference has been proven effective through multiple experiments. This architecture ranks highly on various benchmarks for stereo matching tasks.

2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021 (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Long-range Attention Network for Multi-View Stereo

Xudong Zhang et al.

Summary: LANet introduces a Long-range Attention Network to capture long-range interdependence across the entire space, and proposes a new loss to supervise the distribution of the intermediate probability volume. Extensive experiments on large-scale DTU dataset demonstrate that LANet outperforms previous methods.

2021 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION WACV 2021 (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Cost Volume Pyramid Based Depth Inference for Multi-View Stereo

Jiayu Yang et al.

2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) (2020)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Fast-MVSNet: Sparse-to-Dense Multi-View Stereo With Learned Propagation and Gauss-Newton Refinement

Zehao Yu et al.

2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) (2020)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

DeepPruner: Learning Efficient Stereo Matching via Differentiable PatchMatch

Shivam Duggal et al.

2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019) (2019)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Leveraging Heterogeneous Auxiliary Tasks to Assist Crowd Counting

Muming Zhao et al.

2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019) (2019)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Multi-Scale Geometric Consistency Guided Multi-View Stereo

Qingshan Xu et al.

2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019) (2019)

Add to Collection

Article Computer Science, Software Engineering

Tanks and Temples: Benchmarking Large-Scale Scene Reconstruction

Arno Knapitsch et al.

ACM TRANSACTIONS ON GRAPHICS (2017)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

A New Representation of Skeleton Sequences for 3D Action Recognition

Qiuhong Ke et al.

30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017) (2017)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Unsupervised Monocular Depth Estimation with Left-Right Consistency

Clement Godard et al.

30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017) (2017)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Deformable Convolutional Networks

Jifeng Dai et al.

2017 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV) (2017)

Add to Collection

Article Computer Science, Artificial Intelligence

Large-Scale Data for Multiple-View Stereopsis

Henrik Aanaes et al.

INTERNATIONAL JOURNAL OF COMPUTER VISION (2016)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Massively Parallel Multiview Stereopsis by Surface Normal Diffusion

Silvano Galliani et al.

2015 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV) (2015)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

FlowNet: Learning Optical Flow with Convolutional Networks

Alexey Dosovitskiy et al.

2015 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV) (2015)

Add to Collection

Article Computer Science, Artificial Intelligence

Efficient large-scale multi-view stereo for ultra high-resolution image sets

Engin Tola et al.

MACHINE VISION AND APPLICATIONS (2012)

Add to Collection

Article Computer Science, Artificial Intelligence

Accurate, Dense, and Robust Multiview Stereopsis

Yasutaka Furukawa et al.

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE (2010)

Add to Collection

© Peeref 2019-2024. All rights reserved.