Density-aware Haze Image Synthesis by Self-Supervised Content-Style Disentanglement
This work addresses the incremental improvement of haze image synthesis for computer vision applications, specifically enhancing the generalization of vehicle detectors in hazy conditions.
The paper tackled the problem of incomplete content-style disentanglement in haze image synthesis, which previously led to ill-rendered images, by proposing a self-supervised style regression method that improved disentangling completeness and enabled level-aware haze synthesis, with generated data showing a linear correlation between haze-level and vehicle detection performance degradation.
The key procedure of haze image translation through adversarial training lies in the disentanglement between the feature only involved in haze synthesis, i.e.style feature, and the feature representing the invariant semantic content, i.e. content feature. Previous methods separate content feature apart by utilizing it to classify haze image during the training process. However, in this paper we recognize the incompleteness of the content-style disentanglement in such technical routine. The flawed style feature entangled with content information inevitably leads the ill-rendering of the haze images. To address, we propose a self-supervised style regression via stochastic linear interpolation to reduce the content information in style feature. The ablative experiments demonstrate the disentangling completeness and its superiority in level-aware haze image synthesis. Moreover, the generated haze data are applied in the testing generalization of vehicle detectors. Further study between haze-level and detection performance shows that haze has obvious impact on the generalization of the vehicle detectors and such performance degrading level is linearly correlated to the haze-level, which, in turn, validates the effectiveness of the proposed method.