# Learning deep illumination-robust features from multispectral filter array images

arXiv: https://arxiv.org/abs/2407.15472 · HTML: https://mlsanity.com/paper/2407.15472/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 35 references that are in the map.
- First posted on arXiv 22 Jul 2024.

## Builds on (most cited first, 7 of 7)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [SpectralFormer: Rethinking Hyperspectral Image Classification with Transformers](https://mlsanity.com/paper/2107.02988/) (2021)
- [Convolutional Color Constancy](https://mlsanity.com/paper/1507.00410/) (2015)
- [Learning Raw Image Denoising with Bayer Pattern Unification and Bayer Preserving Augmentation](https://mlsanity.com/paper/1904.12945/) (2019)
- [Multispectral snapshot demosaicing via non-convex matrix completion](https://mlsanity.com/paper/1902.11032/) (2019)
- [Vision Conformer: Incorporating Convolutions into Vision Transformer Layers](https://mlsanity.com/paper/2304.13991/) (2023)

## Abstract (arXiv)

Multispectral (MS) snapshot cameras equipped with a MS filter array (MSFA), capture multiple spectral bands in a single shot, resulting in a raw mosaic image where each pixel holds only one channel value. The fully-defined MS image is estimated from the raw one through textitdemosaicing, which inevitably introduces spatio-spectral artifacts. Moreover, training on fully-defined MS images can be computationally intensive, particularly with deep neural networks (DNNs), and may result in features lacking discrimination power due to suboptimal learning of spatio-spectral interactions. Furthermore, outdoor MS image acquisition occurs under varying lighting conditions, leading to illumination-dependent features. This paper presents an original approach to learn discriminant and illumination-robust features directly from raw images. It involves: textitraw spectral constancy to mitigate the impact of illumination, textitMSFA-preserving transformations suited for raw image augmentation to train DNNs on diverse raw textures, and textitraw-mixing to capture discriminant spatio-spectral interactions in raw images. Experiments on MS image classification show that our approach outperforms both handcrafted and recent deep learning-based methods, while also requiring significantly less computational effort. The source code is available at https://github.com/AnisAmziane/RawTexture.