A Hybrid Dropout Method for High-Precision Seafloor Topography Reconstruction and Uncertainty Quantification
Source abstract
Abstract
Seafloor topography super-resolution reconstruction is critical for marine resource exploration, geological monitoring, and navigation safety. However, sparse acoustic data frequently result in the loss of high-frequency details, and traditional deep learning models exhibit limitations in uncertainty quantification, impeding their practical application. To address these challenges, this study systematically investigates the combined effects of various regularization strategies and uncertainty quantification modules. It proposes a hybrid dropout model that jointly optimizes high-precision reconstruction and uncertainty estimation. The model integrates residual blocks, squeeze-and-excitation (SE) modules, and a multi-scale feature extraction network while employing Monte Carlo Dropout (MC-Dropout) alongside heteroscedastic noise modeling to dynamically gate the uncertainty quantification process. By adaptively modulating the regularization strength based on feature activations, the model preserves high-frequency information and accurately estimates predictive uncertainty. The experimental results demonstrate significant improvements in the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Peak Signal-to-Noise Ratio (PSNR). Compared to conventional dropout architectures, the proposed method achieves a PSNR increase of 46.5% to 60.5% in test regions with a marked reduction in artifacts. Overall, the synergistic effect of employed regularization strategies and uncertainty quantification modules substantially enhances detail recovery and robustness in complex seafloor topography reconstruction, offering valuable theoretical insights and practical guidance for further optimization of deep learning models in challenging applications.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Reservoir Engineering and Simulation Methods · Seismic Imaging and Inversion Techniques · Underwater Acoustics Research
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
School of Electrical Engineering, Naval University of Engineering, Wuhan 430033, China
Key Laboratory of Geological Exploration and Evaluation, Ministry of Education, China University of Geosciences, Wuhan 430074, China
Funding
National Natural Science Foundation of China
References (59)
Show source-provided references
1. Applying Multibeam Sonar and Mathematical Modeling for Mapping Seabed Substrate and Biota of Offshore Shallows10.1016/j.ecss.2017.04.026 ↗
2. On the Use of Digital Elevation Model Data for Hortonian and Fractal Analyses of Channel Networks10.1029/93WR00545 ↗
3. Physically-Based Landslide Prediction over a Large Region: Scaling Low-Resolution Hydrological Model Results for High-Resolution Slope Stability Assessment10.1016/j.envsoft.2019.104607 ↗
4. Lurton, X. (2010). An Introduction to Underwater Acoustics: Principles and Applications, Springer. [2nd ed.].10.1007/978-3-642-13835-5 ↗
5. Seafloor Habitat Mapping Using Multibeam Bathymetric and Backscatter Intensity Multi-Features SVM Classification Framework10.1016/j.apacoust.2020.107728 ↗
6. Long-Term Spatiotemporal Mapping in Lacustrine Environment by Remote Sensing: Review with Case Study, Challenges, and Future Directions10.1016/j.watres.2024.122457 ↗
7. Accuracy Assessment of LiDAR-Derived Digital Elevation Models Based on Approximation Theory10.3390/rs70607062 ↗
8. Mapping Monthly Precipitation, Temperature, and Solar Radiation for Ireland with Polynomial Regression and a Digital Elevation Model10.3354/cr010035 ↗
9. Effects of Terrain Morphology, Sampling Density, and Interpolation Methods on Grid DEM Accuracy10.14358/PERS.71.7.805 ↗
10. Methods for Interpolating Stream Width, Depth, and Current Velocity10.1016/j.ecolmodel.2006.02.004 ↗
11. Fast B-Spline Transforms for Continuous Image Representation and Interpolation10.1109/34.75515 ↗
12. Super-Resolution Reconstruction of a Digital Elevation Model Based on a Deep Residual Network10.1515/geo-2020-0207 ↗
13. Terrain Feature-Aware Deep Learning Network for Digital Elevation Model Superresolution10.1016/j.isprsjprs.2022.04.028 ↗
14. Image Super-Resolution Using Deep Convolutional Networks10.1109/TPAMI.2015.2439281 ↗
15. Kim, J., Lee, J.K., and Lee, K.M. (2016, January 27–30). Accurate Image Super-Resolution Using Very Deep Convolutional Networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA.10.1109/CVPR.2016.182 ↗
16. Ledig, C., Theis, L., Huszár, F., Caballero, J., Cunningham, A., Acosta, A., Aitken, A.P., Tejani, A., Totz, J., and Wang, Z. (2017, January 21–26). Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network. Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA.10.1109/CVPR.2017.19 ↗
17. Underwater Image Super-Resolution via Multi-Scale Feature Fusion
18. Wang, X., Zhang, C., and Zhang, Y. (2023, January 17–24). Multi-Scale Generative Adversarial Network for Underwater Image Restoration. Proceedings of the CVPR Underwater Vision Workshop, Vancouver, BC, Canada.
19. Kendall, A., and Gal, Y. (2017, January 4–9). What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?. Proceedings of the Advances in Neural Information Processing Systems 30, Long Beach, CA, USA.
20. Diverse Image Super-Resolution via Probabilistic Modeling
21. High-Resolution Digital Bathymetric Model (DBM) with Uncertainty Reconstructed by Deep Learning with Dropout
22. Gal, Y., and Ghahramani, Z. (2016, January 20–22). Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning. Proceedings of the 33rd International Conference on Machine Learning, New York, NY, USA.
23. Gal, Y., Hron, J., and Kendall, A. (2017). Concrete Dropout. arXiv.
24. Kingma, D.P., Salimans, T., and Welling, M. (2015, January 7–12). Variational Dropout and the Local Reparameterization Trick. Proceedings of the Advances in Neural Information Processing Systems 28, Montreal, QC, Canada.
25. Louizos, C., Welling, M., and Kingma, D.P. (2017). Learning Sparse Neural Networks through L0 Regularization. arXiv.
26. Molchanov, D., Ashukha, A., and Vetrov, D.P. (2017, January 6–11). Variational Dropout Sparsifies Deep Neural Networks. Proceedings of the 34th International Conference on Machine Learning, Sydney, Australia.
27. Ghiasi, G., Lin, T., and Le, Q.V. (2018, January 3–8). DropBlock: A Regularization Method for Convolutional Networks. Proceedings of the Advances in Neural Information Processing Systems 31, Montreal, QC, Canada. arXiv:1810.12890.
28. Fabbri, C., Islam, M.J., and Sattar, J. (2018, January 28–31). Enhancing Underwater Imagery Using Generative Adversarial Networks. Proceedings of the OCEANS 2018 MTS/IEEE Kobe Techno-Ocean, Kobe, Japan.10.1109/ICRA.2018.8460552 ↗
29. Lakshminarayanan, B., Pritzel, A., and Blundell, C. (2017, January 4–9). Simple and Scalable Predictive Uncertainty Estimation Using Deep Ensembles. Proceedings of the Advances in Neural Information Processing Systems 30, Long Beach, CA, USA. arXiv:1612.01474.
30. Li, Y., Swersky, K., and Zemel, R. (2018, January 10–15). Dropout Inference in Bayesian Neural Networks with Alpha-Divergences. Proceedings of the 35th International Conference on Machine Learning, Stockholm, Sweden. arXiv:1802.02538.
31. Lee, J., and Cho, S. (2018). Bayesian Uncertainty Estimation for Batch Normalized Deep Networks with MC Dropout. arXiv.
32. Frick, T., Antognini, D., Giurgiu, I., Grewe, B.F., Malossi, C., Zhu, R.J.B., and Rigotti, M. (2024). MC Layer Normalization for Calibrated Uncertainty in Deep Learning. Trans. Mach. Learn. Res.
33. Zhang, R., Li, Z., and Sun, H. (2018). Image Super-Resolution Using Very Deep Residual Channel Attention Networks. arXiv.10.1007/978-3-030-01234-2_18 ↗
34. An Enhanced Double-Filter Deep Residual Neural Network for Generating Super Resolution DEMs
35. Spatial Interpolation Using Conditional Generative Adversarial Neural Networks10.1080/13658816.2019.1599122 ↗
36. Hu, J., Shen, L., and Sun, G. (2018, January 18–23). Squeeze-and-Excitation Networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Salt Lake City, UT, USA.10.1109/CVPR.2018.00745 ↗
37. He, K., Zhang, X., Ren, S., and Sun, J. (2016, January 30). Deep Residual Learning for Image Recognition. Proceedings of the Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA.10.1109/CVPR.2016.90 ↗
38. Shi, W., Caballero, J., Huszár, F., Totz, J., Aitken, A.P., Bishop, R., Rueckert, D., and Wang, Z. (2016, January 27–30). Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA.10.1109/CVPR.2016.207 ↗
39. Cubic Convolution Interpolation for Digital Image Processing10.1109/TASSP.1981.1163711 ↗
40. Kong, X., Liu, X., Gu, J., Qiao, Y., and Dong, C. (2022, January 18–24). Reflash Dropout in Image Super-Resolution. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA.10.1109/CVPR52688.2022.00591 ↗
41. Adaptive Sparse Dropout: Learning the Certainty and Uncertainty in Deep Neural Networks10.1016/j.neucom.2021.04.047 ↗
42. Dusenberry, M., Jerfel, G., Wen, Y., Ma, Y., Snoek, J., Heller, K., Lakshminarayanan, B., and Tran, D. (2020, January 12–18). Efficient and Scalable Bayesian Neural Nets with Rank-1 Factors. Proceedings of the 37th International Conference on Machine Learning, Online.
43. Self-Attentive Spatial Adaptive Normalization for Cross-Modality Domain Adaptation10.1109/TMI.2021.3059265 ↗
44. Ioffe, S., and Szegedy, C. (2015). Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. arXiv.
45. Dynamic Neural Networks: A Survey10.1109/TPAMI.2021.3117837 ↗
46. Cho, K., Merrienboer, B.V., Gülçehre, Ç., Bahdanau, D., Bougares, F., Schwenk, H., and Bengio, Y. (2014, January 25–29). Learning Phrase Representations Using RNN Encoder–Decoder for Statistical Machine Translation. Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), Doha, Qatar.10.3115/v1/D14-1179 ↗
47. Smith, L., and Gal, Y. (2018). Understanding Measures of Uncertainty for Adversarial Example Detection. arXiv.
48. Nix, D.A., and Weigend, A.S. (July, January 28). Estimating the Mean and Variance of the Target Probability Distribution. Proceedings of the 1994 IEEE International Conference on Neural Networks (ICNN’94), Orlando, FL, USA.10.1109/ICNN.1994.374138 ↗
49. Image Quality Assessment: From Error Visibility to Structural Similarity10.1109/TIP.2003.819861 ↗
50. Amante, C., and Eakins, B.W. (2009). ETOPO1 1 Arc-Minute Global Relief Model: Procedures, Data Sources and Analysis, NOAA, National Geophysical Data Center, Marine Geology and Geophysics Division Boulder. Technical Report.
51. Global Bathymetry and Topography at 15 Arc Sec: SRTM15+10.1029/2019EA000658 ↗
52. Agarwal, S. (2013, January 21–23). Data Mining: Data Mining Concepts and Techniques. Proceedings of the 2013 International Conference on Machine Intelligence and Research Advancement, Katra, India.10.1109/ICMIRA.2013.45 ↗
53. Guo, C., Pleiss, G., Sun, Y., and Weinberger, K.Q. (2017). On Calibration of Modern Neural Networks. arXiv.
54. Dropout: A Simple Way to Prevent Neural Networks from Overfitting
55. Blundell, C., Cornebise, J., Kavukcuoglu, K., and Wierstra, D. (2015). Weight Uncertainty in Neural Networks. arXiv.
56. Ovadia, Y., Fertig, E., Ren, J., Nado, Z., Sculley, D., Nowozin, S., Dillon, J., Lakshminarayanan, B., and Snoek, J. (2019, January 8–14). Can You Trust Your Model’s Uncertainty? Evaluating Predictive Uncertainty under Dataset Shift. Proceedings of the Advances in Neural Information Processing Systems 32, Vancouver, BC, Canada.
57. Feature-Enhanced Deep Learning Network for Digital Elevation Model Super-Resolution10.1109/JSTARS.2023.3288296 ↗
58. Lin, X., Zhang, Q., Wang, H., Yao, C., Chen, C., Cheng, L., and Li, Z. (2022). A DEM Super-Resolution Reconstruction Network Combining Internal and External Learning. Remote Sens., 14.10.3390/rs14092181 ↗
59. Zhang, R., Bian, S., and Li, H. (2021). RSPCN: Super-Resolution of Digital Elevation Model Based on Recursive Sub-Pixel Convolutional Neural Networks. ISPRS Int. J. Geo-Inf., 10.10.3390/ijgi10080501 ↗