Multispectral Image Reconstruction From Color Images Using Enhanced Variational Autoencoder and Generative Adversarial Network

• Published in: IEEE access Vol. 9; pp. 1666 - 1679
• Main Authors: Liu, Xu, Gherbi, Abdelouahed, Wei, Zhenzhou, Li, Wubin, Cheriet, Mohamed
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Color imagery; color vision; Datasets; Gallium nitride; Generative adversarial network (GAN); Generative adversarial networks; Generators; Hyperspectral imaging; Image color analysis; Image enhancement; image processing; Image reconstruction; multispectral image; Neural networks; Normal distribution; RGB image; spectral reconstruction; stochastic neural network; VAE-GAN; variational autoencoder (VAE)
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2020.3047074

Since multispectral images (MSIs) have much more sufficient spectral information than RGB images (RGBs), reconstructing MS images from RGB images is a severely underconstrained problem. We have to generate colossally different information between the two scopes. Almost all previous approaches are based on static and dependent neural networks, which fail to explain how to supplement the massive lost information. This paper presents a low-cost and high-efficiency approach, "VAE-GAN", based on stochastic neural networks to directly reconstruct high-quality MSIs from RGBs. Our approach combines the advantages of the Generative Adversarial Network (GAN) and the Variational Autoencoder (VAE). The VAE undertakes the generation of the lost variational MS distributions by reparameterizing the latent space vector with sampling from Gaussian distribution. The GAN is responsible for regulating the generator to produce MSI-like images. In this way, our approach can create huge missed information and make the outputs look real, which also solves the previous problem. Moreover, we use several qualitative and quantitative methods to evaluate our approach and obtain excellent results. In particular, with much less training data than the previous approaches, we obtained comparable results on the CAVE dataset and surpassed state-of-the-art results on the ICVL dataset.