Joint haze image synthesis and dehazing with mmd-vae losses
This work addresses perception enhancement and testing for autonomous vehicles, but it is incremental as it builds on existing VAE and GAN methods with a specific modification.
The paper tackles the dual problems of haze image synthesis and dehazing for autonomous driving systems by proposing an unsupervised Image-to-Image Translation framework using VAE and GAN with MMD losses to handle imbalanced and unpaired training data, demonstrating feasibility and practicability in both tasks.
Fog and haze are weathers with low visibility which are adversarial to the driving safety of intelligent vehicles equipped with optical sensors like cameras and LiDARs. Therefore image dehazing for perception enhancement and haze image synthesis for testing perception abilities are equivalently important in the development of such autonomous driving systems. From the view of image translation, these two problems are essentially dual with each other, which have the potentiality to be solved jointly. In this paper, we propose an unsupervised Image-to-Image Translation framework based on Variational Autoencoders (VAE) and Generative Adversarial Nets (GAN) to handle haze image synthesis and haze removal simultaneously. Since the KL divergence in the VAE objectives could not guarantee the optimal mapping under imbalanced and unpaired training samples with limited size, Maximum mean discrepancy (MMD) based VAE is utilized to ensure the translating consistency in both directions. The comprehensive analysis on both synthesis and dehazing performance of our method demonstrate the feasibility and practicability of the proposed method.