Group-Wise Correlation Stereo Network

By Xiaoyang Guo, Kai Yang, Wukui Yang, Xiaogang Wang, Hongsheng Li. 2019

718 citations recorded by the source corpus.

Stereo matching estimates the disparity between a rectified image pair, which is of great importance to depth sensing, autonomous driving, and other related tasks. Previous works built cost volumes with cross-correlation or concatenation of left and right features across all disparity levels, and then a 2D or 3D convolutional neural network is utilized to regress the disparity maps. In this paper, we propose to construct the cost volume by group-wise correlation. The left features and the right features are divided into groups along the channel dimension, and correlation maps are computed among each group to obtain multiple matching cost proposals, which are then packed into a cost volume. Group-wise correlation provides efficient representations for measuring feature similarities and will not lose too much information like full correlation. It also preserves better performance when reducing parameters compared with previous methods. The 3D stacked hourglass network proposed in previous works is improved to boost the performance and decrease the inference computational cost. Experiment results show that our method outperforms previous methods on Scene Flow, KITTI 2012, and KITTI 2015 datasets.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

IRS: A Large Synthetic Indoor Robotics Stereo Dataset for Disparity and Surface Normal Estimation (2019, background): “Recent advances [1], [2], [3], [4], [5], [6] in these vision tasks have shown that deep neural network (DNN) can significantly improve the estimation quality.”
CFNet: Cascade and Fused Cost Volume for Robust Stereo Matching (2021, methodology): “Cost volume construction: Inspired by [26, 12], we propose to use both feature concatenation and group-wise correlation to generate combination volume.”
Practical Stereo Matching via Cascaded Recurrent Network with Adaptive Correlation (2022, methodology): “One line of networks [12, 21, 22, 27, 29, 44, 49] only uses 2D convolutions.”
Attention Concatenation Volume for Accurate and Efficient Stereo Matching (2022, methodology): “To overcome the above drawbacks, GwcNet [7] proposes the group-wise correlation volume and concatenates it with a compact concatenation volume to form a combined volume, which aims to combine the advantages of two volumes.”
AANet: Adaptive Aggregation Network for Efficient Stereo Matching (2020, background): “An exception to concatenation based 3D methods is [8], in which group-wise correlation is proposed to reduce the information loss of full correlation [20].”
RAFT-Stereo: Multilevel Recurrent Field Transforms for Stereo Matching (2021, methodology): “Next, a 3D cost volume is constructed over the 2D feature maps, either through concatenation[19] or correlation operator[12].”
Iterative Geometry Encoding Volume for Stereo Matching (2023, methodology): “GwcNet [17] and ACVNet [47] propose the group-wise correlation volume and the attention concatenation volume, respectively, to improve the expressiveness of the cost volume and in turn improve the performance in ambiguous regions.”
Cascade Cost Volume for High-Resolution Multi-View Stereo and Stereo Matching (2019, methodology): “Typically, these methods downsample the feature maps to formulate the cost volumes at a lower resolution [3,4,8,12,15,22,26,38,44,45,48], and adopt upsampling [3, 8, 15, 26, 34, 38, 41, 48] or postrefinement [4, 22] to output the final high-resolution result.”
PatchmatchNet: Learned Multi-View Patchmatch Stereo (2020, methodology): “Then we apply group-wise correlation [18] be-tween the features at each sampling location and p .”
Deep learning in optical metrology: a review (2022, methodology): “On a different note, Guo et al.344 improved the 3D-stacked hourglass network to obtain the cost volume by group-wise correlation and then realized stereomatching.”
A Survey on Deep Learning Techniques for Stereo-Based Depth Estimation (2020, methodology): “To beneﬁt from both, Guo et al. [86] propose a hybrid approach, which constructs two cost volumes; one by feature concatenation but compressed into 12 channels using two convolutions.”
Real-Time Self-Adaptive Deep Stereo (2018, methodology): “Thereby, [21] ignited the spread of end-to-end stereo architectures [17, 24, 19, 4, 16, 11] that quickly outmatched any other technique on the KITTI benchmarks by leveraging on a peculiar training protocol.”
Revisiting Stereo Depth Estimation From a Sequence-to-Sequence Perspective with Transformers (2020, methodology): “Under multiple evaluation settings, we compare STTR with prior work spanning the primary learning-based stereo depth paradigms, including correlation-based AANet [39], 3D convolution-based PSMNet [5] and GANet-11 [43], correlation and 3D convolution hybrid approach GwcNet-g [16] and a classification”