Zhenzhou Wei

Ms. Zhenzhou Wei


Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
24 December 2020 in IEEE Access
Verified
Journal Article
Open access
XLXu Liu
Xu Liu
AGAbdelouahed Gherbi
Abdelouahed Gherbi
ZWZhenzhou Wei
Zhenzhou Wei
WLWubin Li
Wubin Li
MCMohamed Cheriet
Mohamed Cheriet
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.

10 August 2021 in Sensors
Verified
Journal Article
Open access
PDF file
XLXu Liu
Xu Liu
AGAbdelouahed Gherbi
Abdelouahed Gherbi
WLWubin Li
Wubin Li
ZWZhenzhou Wei
Zhenzhou Wei
MCMohamed Cheriet
Mohamed Cheriet
Since multispectral images (MSIs) and RGB images (RGBs) have significantly different definitions and severely imbalanced information entropies, the spectrum transformation between them, especially reconstructing MSIs from RGBs, is a big challenge. We propose a new approach, the Taiji Generative Neural Network (TaijiGNN), to address the above-mentioned problems. TaijiGNN consists of two generators, G_MSI, and G_RGB. These two generators establish two cycles by connecting one generator’s output with the other’s input. One cycle translates the RGBs into the MSIs and converts the MSIs back to the RGBs. The other cycle does the reverse. The cycles can turn the problem of comparing two different domain images into comparing the same domain images. In the same domain, there are neither different domain definition problems nor severely underconstrained challenges, such as reconstructing MSIs from RGBs. Moreover, according to several investigations and validations, we effectively designed a multilayer perceptron neural network (MLP) to substitute the convolutional neural network (CNN) when implementing the generators to make them simple and high performance. Furthermore, we cut off the two traditional CycleGAN’s identity losses to fit the spectral image translation. We also added two consistent losses of comparing paired images to improve the two generators’ training effectiveness. In addition, during the training process, similar to the ancient Chinese philosophy Taiji’s polarity Yang and polarity Yin, the two generators update their neural network parameters by interacting with and complementing each other until they all converge and the system reaches a dynamic balance. Furthermore, several qualitative and quantitative experiments were conducted on the two classical datasets, CAVE and ICVL, to evaluate the performance of our proposed approach. Promising results were obtained with a well-designed simplistic MLP requiring a minimal amount of training data. Specifically, in the CAVE dataset, to achieve comparable state-of-the-art results, we only need half of the dataset for training; for the ICVL dataset, we used only one-fifth of the dataset to train the model, but obtained state-of-the-art results.