Papers
Topics
Authors
Recent
Search
2000 character limit reached

Cutting-Edge Techniques for Depth Map Super-Resolution

Published 27 Jun 2023 in cs.CV and eess.IV | (2306.15244v1)

Abstract: To overcome hardware limitations in commercially available depth sensors which result in low-resolution depth maps, depth map super-resolution (DMSR) is a practical and valuable computer vision task. DMSR requires upscaling a low-resolution (LR) depth map into a high-resolution (HR) space. Joint image filtering for DMSR has been applied using spatially-invariant and spatially-variant convolutional neural network (CNN) approaches. In this project, we propose a novel joint image filtering DMSR algorithm using a Swin transformer architecture. Furthermore, we introduce a Nonlinear Activation Free (NAF) network based on a conventional CNN model used in cutting-edge image restoration applications and compare the performance of the techniques. The proposed algorithms are validated through numerical studies and visual examples demonstrating improvements to state-of-the-art performance while maintaining competitive computation time for noisy depth map super-resolution.

Authors (2)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (22)
  1. The fast bilateral solver. In Proc. ECCV, pages 617–632. Springer, 2016.
  2. Swin-UNet: UNet-like pure transformer for medical image segmentation. arXiv preprint arXiv:2105.05537, 2021.
  3. Simple baselines for image restoration. arXiv preprint arXiv:2204.04676, Apr. 2022.
  4. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020.
  5. Deep sparse rectifier neural networks. In Proc. 14th Int. Conf. Artif. Intell. Stat. (ISTATS), pages 315–323, Fort Lauderdale, FL, USA, Apr. 2011.
  6. Robust guided image filtering using nonconvex potentials. IEEE Trans. Pattern Anal. Mach. Intell., 40(1):192–207, 2017.
  7. Towards fast and accurate real-world depth super-resolution: Benchmark dataset and baseline. In Proc. IEEE CVPR, pages 9229–9238, 2021.
  8. Gaussian error linear units (GELUs). arXiv preprint arXiv:1606.08415, 2016.
  9. Fast cost-volume filtering for visual correspondence and beyond. IEEE Trans. Pattern Anal. Mach. Intell., 35(2):504–511, 2012.
  10. Squeeze-and-excitation networks. In Proc. IEEE CVPR, pages 7132–7141, Salt Lake City, Utah, USA, June 2018.
  11. Deformable kernel networks for joint image filtering. IJCV, 129(2):579–600, 2021.
  12. Deep joint image filtering. In Proc. ECCV, pages 154–169, 2016.
  13. Swinir: Image restoration using swin transformer. In Proc. IEEE ICCV, pages 1833–1844, 2021.
  14. Swin transformer v2: Scaling up capacity and resolution. arXiv preprint arXiv:2111.09883, 2021.
  15. Swin transformer: Hierarchical vision transformer using shifted windows. In Proc. IEEE ICCV, pages 10012–10022, 2021.
  16. Mutual-structure for joint filtering. In Proc. IEEE CVPR, pages 3406–3414, 2015.
  17. Indoor segmentation and support inference from RGB-D images. In Proc. ECCV, pages 746–760. Springer, 2012.
  18. A vision transformer approach for efficient near-field SAR super-resolution under array perturbation. In IEEE Proc. TSWMCS, Waco, TX, USA, Apr. 2022.
  19. Attention is all you need. Proc. NeurIPS, 30, 2017.
  20. Self-supervised learning with swin transformers. arXiv preprint arXiv:2105.04553, 2021.
  21. Rolling guidance filter. In Proc. ECCV, pages 815–830, 2014.
  22. High-resolution depth maps imaging via attention-based hierarchical multi-modal fusion. IEEE Trans. Image Process., 31:648–663, 2021.

Summary

No one has generated a summary of this paper yet.

Sign Up to Summarize

Paper to Video (Beta)

No one has generated a video about this paper yet.

Sign Up to Generate All Videos Subscribe on YouTube

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Sign Up to Generate

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Generate Now

Continue Learning

We haven't generated follow-up questions for this paper yet.

Generate Now

Collections

Sign up for free to add this paper to one or more collections.

Sign Up