Multi-Level Context Ultra-Aggregation for Stereo Matching

Nie, Guang-Yu; Cheng, Ming–Ming; Liu, Yun; Liang, Zhengfa; Fan, Deng-Ping; Liu, Yue; Wang, Yongtian

doi:10.1109/cvpr.2019.00340

Cited by 111 publications

(56 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Following this seminal work, [13] proposed an end-to-end architecture inspired by the classical stereo matching pipeline. Many works followed this trend [2,23,20,18,24,9,5] and completely outmatched previous methods in terms of accuracy. However, this gain in accuracy is at the cost of high computation requirement.…”

Section: Related Workmentioning

confidence: 95%

Fast Stereo Disparity Maps Refinement By Fusion of Data-Based And Model-Based Estimations

Ferrera

Boulch

Moras

2019

2019 International Conference on 3D Vision (3DV)

View full text Add to dashboard Cite

The estimation of disparity maps from stereo pairs has many applications in robotics and autonomous driving. Stereo matching has first been solved using model-based approaches, with real-time considerations for some, but today's most recent works rely on deep convolutional neural networks and mainly focus on accuracy at the expense of computing time. In this paper, we present a new method for disparity maps estimation getting the best of both worlds: the accuracy of data-based methods and the speed of fast model-based ones. The proposed approach fuses prior disparity maps to estimate a refined version. The core of this fusion pipeline is a convolutional neural network that leverages dilated convolutions for fast context aggregation without spatial resolution loss. The resulting architecture is both very effective for the task of refining and fusing prior disparity maps and very light, allowing our fusion pipeline to produce disparity maps at rates up to 125 Hz. We obtain state-of-the-art results in terms of speed and accuracy on the KITTI benchmarks. Code and pre-trained models are available on our github: https://github.com/ ferreram/FD-Fusion.

show abstract

Section: Related Workmentioning

confidence: 95%

Fast Stereo Disparity Maps Refinement By Fusion of Data-Based And Model-Based Estimations

Ferrera

Boulch

Moras

2019

2019 International Conference on 3D Vision (3DV)

View full text Add to dashboard Cite

show abstract

“…Existing end-to-end disparity estimation networks usually include cost volume computation, cost aggregation, and disparity prediction. 2D CNN based methods [17,20,33] generally adopt a correlation layer for 3D cost volume construction, while 3D CNN based methods [2,3,8,22,38] mostly use direct feature concatenation to construct 4D cost volume and use 3D convolution for cost aggregation. Apart from supervised methods, several unsupervised learning methods [14,24,29,36,40] have been developed to avoid the use of costly ground truth depth annotations.…”

Section: Disparity Estimationmentioning

confidence: 99%

Feedback Network for Mutually Boosted Stereo Image Super-Resolution and Disparity Estimation

Dai

et al. 2021

Proceedings of the 29th ACM International Conference on Multimedia

View full text Add to dashboard Cite

Under stereo settings, the problem of image super-resolution (SR) and disparity estimation are interrelated that the result of each problem could help to solve the other. The effective exploitation of correspondence between different views facilitates the SR performance, while the high-resolution (HR) features with richer details benefit the correspondence estimation. According to this motivation, we propose a Stereo Super-Resolution and Disparity Estimation Feedback Network (SSRDE-FNet), which simultaneously handles the stereo image super-resolution and disparity estimation in a unified framework and interact them with each other to further improve their performance. Specifically, the SSRDE-FNet is composed of two dual recursive sub-networks for left and right views. Besides the cross-view information exploitation in the low-resolution (LR) space, HR representations produced by the SR process are utilized to perform HR disparity estimation with higher accuracy, through which the HR features can be aggregated to generate a finer SR result. Afterward, the proposed HR Disparity Information Feedback (HRDIF) mechanism delivers information carried by HR disparity back to previous layers to further refine the SR image reconstruction. Extensive experiments demonstrate the effectiveness and advancement of SSRDE-FNet.

show abstract

“…It plays a key role in a range of real-world applications, such as stereo matching [2], image understanding [3], co-saliency detection [4], action recognition [5], video detection and segmentation [6][7][8][9], semantic segmentation [10,11], medical image segmentation [12][13][14], object tracking [15,16], person re-identification [17,18], camouflaged object detection [19], image retrieval [20], etc. Although significant progress has been made in the salient object detection field over the past several years [21][22][23][24][25][26][27][28][29][30][31][32][33][34][35], there is still room for improvement when faced with challenging factors, such as complicated backgrounds or varying lighting conditions in the scenes.…”

Section: Introductionmentioning

confidence: 99%

“…(2) According to fusion model: it is critical to effectively fuse RGB and depth images in this task, so we review different fusion strategies to understand their effectiveness. (3) As single-or multi-stream models: using a single stream can reduce the number of parameters, but the final result may not be optimal; multiple streams may require more parameters.…”

mentioning

confidence: 99%

RGB-D salient object detection: A survey

et al. 2021

Self Cite

View full text Add to dashboard Cite

Salient object detection, which simulates human visual perception in locating the most significant object(s) in a scene, has been widely applied to various computer vision tasks. Now, the advent of depth sensors means that depth maps can easily be captured; this additional spatial information can boost the performance of salient object detection. Although various RGB-D based salient object detection models with promising performance have been proposed over the past several years, an in-depth understanding of these models and the challenges in this field remains lacking. In this paper, we provide a comprehensive survey of RGB-D based salient object detection models from various perspectives, and review related benchmark datasets in detail. Further, as light fields can also provide depth maps, we review salient object detection models and popular benchmark datasets from this domain too. Moreover, to investigate the ability of existing models to detect salient objects, we have carried out a comprehensive attribute-based evaluation of several representative RGB-D based salient object detection models. Finally, we discuss several challenges and open directions of RGB-D based salient object detection for future research. All collected models, benchmark datasets, datasets constructed for attribute-based evaluation, and related code are publicly available at https://github.com/taozh2017/RGBD-SODsurvey.

show abstract

Multi-Level Context Ultra-Aggregation for Stereo Matching

Cited by 111 publications

References 21 publications

Fast Stereo Disparity Maps Refinement By Fusion of Data-Based And Model-Based Estimations

Fast Stereo Disparity Maps Refinement By Fusion of Data-Based And Model-Based Estimations

Feedback Network for Mutually Boosted Stereo Image Super-Resolution and Disparity Estimation

RGB-D salient object detection: A survey

Contact Info

Product

Resources

About