# Accurate Spectral Super-resolution from Single RGB Image Using Multi-scale CNN

arXiv: https://arxiv.org/abs/1806.03575 · HTML: https://mlsanity.com/paper/1806.03575/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 10 Jun 2018.

## Built on by (most cited first, 9 of 9)

- [A survey on computational spectral reconstruction methods from RGB to hyperspectral imaging](https://mlsanity.com/paper/2106.15944/) (2021)
- [Implicit Neural Representation Learning for Hyperspectral Image Super-Resolution](https://mlsanity.com/paper/2112.10541/) (2021)
- [GMSR:Gradient-Guided Mamba for Spectral Reconstruction from RGB Images](https://mlsanity.com/paper/2405.07777/) (2024)
- [HPRN: Holistic Prior-embedded Relation Network for Spectral Super-Resolution](https://mlsanity.com/paper/2112.14608/) (2021)
- [Deep-learning-based Hyperspectral imaging through a RGB camera](https://mlsanity.com/paper/2107.05190/) (2021)
- [Learning to Recover Spectral Reflectance from RGB Images](https://mlsanity.com/paper/2304.02162/) (2023)
- [LD-GAN: Low-Dimensional Generative Adversarial Network for Spectral Image Generation with Variance Regularization](https://mlsanity.com/paper/2305.00132/) (2023)
- [mHealth hyperspectral learning for instantaneous spatiospectral imaging of hemodynamics](https://mlsanity.com/paper/2303.16205/) (2023)
- [Model-Guided Network with Cluster-Based Operators for Spatio-Spectral Super-Resolution](https://mlsanity.com/paper/2505.24605/) (2025)

## Abstract (arXiv)

Different from traditional hyperspectral super-resolution approaches that focus on improving the spatial resolution, spectral super-resolution aims at producing a high-resolution hyperspectral image from the RGB observation with super-resolution in spectral domain. However, it is challenging to accurately reconstruct a high-dimensional continuous spectrum from three discrete intensity values at each pixel, since too much information is lost during the procedure where the latent hyperspectral image is downsampled (e.g., with x10 scaling factor) in spectral domain to produce an RGB observation. To address this problem, we present a multi-scale deep convolutional neural network (CNN) to explicitly map the input RGB image into a hyperspectral image. Through symmetrically downsampling and upsampling the intermediate feature maps in a cascading paradigm, the local and non-local image information can be jointly encoded for spectral representation, ultimately improving the spectral reconstruction accuracy. Extensive experiments on a large hyperspectral dataset demonstrate the effectiveness of the proposed method.