Subsampled Turbulence Removal Network
This work addresses turbulence removal in videos, which is important for applications like surveillance and astronomy, but it appears incremental as it builds on existing GAN methods with specific enhancements.
The paper tackles the problem of restoring turbulence-distorted video frames by developing a deep-learning approach that uses a novel data augmentation method to model turbulence from small datasets and a subsampling technique to filter out severely corrupted frames. The proposed Wasserstein GAN with ℓ1 cost successfully restores video sequences, achieving improved quality as demonstrated in experiments.
We present a deep-learning approach to restore a sequence of turbulence-distorted video frames from turbulent deformations and space-time varying blurs. Instead of requiring a massive training sample size in deep networks, we purpose a training strategy that is based on a new data augmentation method to model turbulence from a relatively small dataset. Then we introduce a subsampled method to enhance the restoration performance of the presented GAN model. The contributions of the paper is threefold: first, we introduce a simple but effective data augmentation algorithm to model the turbulence in real life for training in the deep network; Second, we firstly purpose the Wasserstein GAN combined with $\ell_1$ cost for successful restoration of turbulence-corrupted video sequence; Third, we combine the subsampling algorithm to filter out strongly corrupted frames to generate a video sequence with better quality.