Autors: Stankovski T., Christoff, N. V., Petkova, R. R., Tonchev K., Manolova, A. H.
Title: From Noisy Data to Realistic Scenes: Neural Network Solutions for XR
Keywords: Denoising CNN, Residual Learning, Visual data enhancement

Abstract: With the development of technologies, in the context of mixed physical and virtual environments (XR) projects, one of the main tasks to be overcome is high-quality visual information for accurate 3D modelling of natural scenes. In this regard, noise and sparse sensor data lead to loss of details and inaccuracies in reconstruction. In this paper, we propose a solution to this problem by developing and integrating an image filtration algorithm based on neural networks. The proposed solution effectively reduces noise and improves the quality of the input visual data, which would serve for further 3D modelling and improvement of XR applications.

References

  1. V. Bhaskaran and U. Mahbub, "Immersive User Experiences: Trends and Challenges of Using XR Technologies", Computer Vision, 260-278, 2024.
  2. B. Zhang and J. P. Allebach, "Adaptive bilateral filter for sharpness enhancement and noise removal", IEEE transactions on Image Processing, 17(5), 664-678, 2008.
  3. W. Li, H et al., "A deep learning method for denoising based on a fast and flexible convolutional neural network", IEEE Trans. on Geo. and Rem. Sen., vol. 60, pp. 1-13, 2021.
  4. Y Li and Z. Ma, "Deep learning-based noise reduction for seismic data", In Journal of Physics: Conference Series, vol. 1861 (1), p. 012011, IOP Publishing, 2021.
  5. T. Taleb et al., "Toward supporting XR services: Architecture and enablers", IEEE Int. of Th. J., vol. 10(4), pp. 3567-3586, 2022.

Issue

60th International Scientific Conference on Information, Communication and Energy Systems and Technologies, ICEST 2025 - Proceedings, 2025, Macedonia, https://doi.org/10.1109/ICEST66328.2025.11098314

Вид: публикация в международен форум, публикация в реферирано издание, индексирана в Scopus