2T-UNET: A Two-Tower UNet with Depth Clues for Robust Stereo Depth Estimation
Not on the map. Why?
Abstract (arXiv)
Stereo correspondence matching is an essential part of the multi-step stereo depth estimation process. This paper revisits the depth estimation problem, avoiding the explicit stereo matching step using a simple two-tower convolutional neural network. The proposed algorithm is entitled as 2T-UNet. The idea behind 2T-UNet is to replace cost volume construction with twin convolution towers. These towers have an allowance for different weights between them. Additionally, the input for twin encoders in 2T-UNet are different compared to the existing stereo methods. Generally, a stereo network takes a right and left image pair as input to determine the scene geometry. However, in the 2T-UNet model, the right stereo image is taken as one input and the left stereo image along with its monocular depth clue information, is taken as the other input. Depth clues provide complementary suggestions that help enhance the quality of predicted scene geometry. The 2T-UNet surpasses state-of-the-art monocular and stereo depth estimation methods on the challenging Scene flow dataset, both quantitatively and qualitatively. The architecture performs incredibly well on complex natural scenes, highlighting its usefulness for various real-time applications. Pretrained weights and code will be made readily available.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 25 of its 43 references that are in the map. First posted on arXiv 27 Oct 2022.
Builds on
25 of 43 references.
References · 25
- 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
- 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
- 2019Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfercited by 1,484
- 2014Depth Map Prediction from a Single Image using a Multi-Scale Deep Networkcited by 1,264
- 2015A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimationcited by 878
- 2020AdaBins: Depth Estimation using Adaptive Binscited by 596
- 2018Pyramid Stereo Matching Networkcited by 486
- 2016Deeper Depth Prediction with Fully Convolutional Residual Networkscited by 402
- 2019GA-Net: Guided Aggregation Net for End-to-end Stereo Matchingcited by 242
- 2020Revisiting Stereo Depth Estimation From a Sequence-to-Sequence Perspective with Transformerscited by 190
- 2018Revisiting Single Image Depth Estimation: Toward Higher Resolution Maps with Accurate Object Boundariescited by 144
- 2018High Quality Monocular Depth Estimation via Transfer Learningcited by 133
13 more
- 20193D Ken Burns Effect from a Single Imagecited by 128
- 2019Self-Supervised Monocular Depth Hintscited by 127
- 2019Hierarchical Deep Stereo Matching on High-resolution Imagescited by 126
- 2019DPSNet: End-to-end Deep Plane Sweep Stereocited by 115
- 2019DeepPruner: Learning Efficient Stereo Matching via Differentiable PatchMatchcited by 114
- 2018Hierarchical Discrete Distribution Decomposition for Match Density Estimationcited by 93
- 2020A Survey on Deep Learning Techniques for Stereo-based Depth Estimationcited by 90
- 2018Real-time self-adaptive deep stereocited by 80
- 2021Channel-Wise Attention-Based Network for Self-Supervised Monocular Depth Estimationcited by 63
- 2018Anytime Stereo Image Depth Estimation on Mobile Devicescited by 54
- 2021A Hierarchical Coding Scheme for Glasses-free 3D Displays Based on Scalable Hybrid Layered Representation of Real-World Light Fieldscited by 2
- 2021A Novel 3D-UNet Deep Learning Framework Based on High-Dimensional Bilateral Grid for Edge Consistent Single Image Depth Estimationcited by 2
- 2020A-TVSNet: Aggregated Two-View Stereo Network for Multi-View Stereo Depth Estimationcited by 1
Built on by
Not cited yet.