Background

RGB-D data image super-resulation NeRF

Model overview Method Scene Representation render color as a weighted sum of radiance values along a ray.[1] Degradation Model High-order degradation Model[2] Sampling Networks and Training

Experiments Datasets and Implementation Comparisons with prior works

Ablation studies

Conclusion

References
[1] Azinović D, Martin-Brualla R, Goldman D B, et al. Neural RGB-D surface reconstruction[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022: 6290-6301.
[2] Wang X, Xie L, Dong C, et al. Real-esrgan: Training real-world blind super-resolution with pure synthetic data[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. 2021: 1905-1914.
[3] Zamir S W, Arora A, Khan S, et al. Restormer: Efficient transformer for high-resolution image restoration[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022: 5728-5739.
[4] Sitzmann V, Martel J, Bergman A, et al. Implicit neural representations with periodic activation functions[J]. Advances in Neural Information Processing Systems, 2020, 33: 7462-7473.